xref: /netbsd-src/sys/netinet/raw_ip.c (revision c38e7cc395b1472a774ff828e46123de44c628e9)
1 /*	$NetBSD: raw_ip.c,v 1.176 2018/04/28 13:26:57 maxv 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 1.176 2018/04/28 13:26:57 maxv Exp $");
69 
70 #ifdef _KERNEL_OPT
71 #include "opt_inet.h"
72 #include "opt_ipsec.h"
73 #include "opt_mrouting.h"
74 #include "opt_net_mpsafe.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 
90 #include <netinet/in.h>
91 #include <netinet/in_systm.h>
92 #include <netinet/ip.h>
93 #include <netinet/ip_var.h>
94 #include <netinet/ip_private.h>
95 #include <netinet/ip_mroute.h>
96 #include <netinet/ip_icmp.h>
97 #include <netinet/in_pcb.h>
98 #include <netinet/in_proto.h>
99 #include <netinet/in_var.h>
100 
101 #ifdef IPSEC
102 #include <netipsec/ipsec.h>
103 #endif
104 
105 struct inpcbtable rawcbtable;
106 
107 int	 rip_pcbnotify(struct inpcbtable *, struct in_addr,
108     struct in_addr, int, int, void (*)(struct inpcb *, int));
109 static int	 rip_connect_pcb(struct inpcb *, struct sockaddr_in *);
110 static void	 rip_disconnect1(struct inpcb *);
111 
112 static void sysctl_net_inet_raw_setup(struct sysctllog **);
113 
114 /*
115  * Nominal space allocated to a raw ip socket.
116  */
117 #define	RIPSNDQ		8192
118 #define	RIPRCVQ		8192
119 
120 static u_long		rip_sendspace = RIPSNDQ;
121 static u_long		rip_recvspace = RIPRCVQ;
122 
123 /*
124  * Raw interface to IP protocol.
125  */
126 
127 /*
128  * Initialize raw connection block q.
129  */
130 void
131 rip_init(void)
132 {
133 
134 	sysctl_net_inet_raw_setup(NULL);
135 	in_pcbinit(&rawcbtable, 1, 1);
136 }
137 
138 static void
139 rip_sbappendaddr(struct inpcb *last, struct ip *ip, const struct sockaddr *sa,
140     int hlen, struct mbuf *n)
141 {
142 	struct mbuf *opts = NULL;
143 
144 	if (last->inp_flags & INP_NOHEADER)
145 		m_adj(n, hlen);
146 	if (last->inp_flags & INP_CONTROLOPTS ||
147 	    SOOPT_TIMESTAMP(last->inp_socket->so_options))
148 		ip_savecontrol(last, &opts, ip, n);
149 	if (sbappendaddr(&last->inp_socket->so_rcv, sa, n, opts) == 0) {
150 		soroverflow(last->inp_socket);
151 		m_freem(n);
152 		if (opts)
153 			m_freem(opts);
154 	} else {
155 		sorwakeup(last->inp_socket);
156 	}
157 }
158 
159 /*
160  * Setup generic address and protocol structures
161  * for raw_input routine, then pass them along with
162  * mbuf chain.
163  */
164 void
165 rip_input(struct mbuf *m, ...)
166 {
167 	int hlen, proto;
168 	struct ip *ip = mtod(m, struct ip *);
169 	struct inpcb_hdr *inph;
170 	struct inpcb *inp;
171 	struct inpcb *last = NULL;
172 	struct mbuf *n;
173 	struct sockaddr_in ripsrc;
174 	va_list ap;
175 
176 	va_start(ap, m);
177 	(void)va_arg(ap, int);		/* ignore value, advance ap */
178 	proto = va_arg(ap, int);
179 	va_end(ap);
180 
181 	sockaddr_in_init(&ripsrc, &ip->ip_src, 0);
182 
183 	/*
184 	 * XXX Compatibility: programs using raw IP expect ip_len
185 	 * XXX to have the header length subtracted, and in host order.
186 	 * XXX ip_off is also expected to be host order.
187 	 */
188 	hlen = ip->ip_hl << 2;
189 	ip->ip_len = ntohs(ip->ip_len) - hlen;
190 	NTOHS(ip->ip_off);
191 
192 	TAILQ_FOREACH(inph, &rawcbtable.inpt_queue, inph_queue) {
193 		inp = (struct inpcb *)inph;
194 		if (inp->inp_af != AF_INET)
195 			continue;
196 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
197 			continue;
198 		if (!in_nullhost(inp->inp_laddr) &&
199 		    !in_hosteq(inp->inp_laddr, ip->ip_dst))
200 			continue;
201 		if (!in_nullhost(inp->inp_faddr) &&
202 		    !in_hosteq(inp->inp_faddr, ip->ip_src))
203 			continue;
204 
205 		if (last == NULL) {
206 			;
207 		}
208 #if defined(IPSEC)
209 		else if (ipsec_used && ipsec_in_reject(m, last)) {
210 			/* do not inject data into pcb */
211 		}
212 #endif
213 		else if ((n = m_copypacket(m, M_DONTWAIT)) != NULL) {
214 			rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, n);
215 		}
216 
217 		last = inp;
218 	}
219 
220 #if defined(IPSEC)
221 	if (ipsec_used && last != NULL && ipsec_in_reject(m, last)) {
222 		m_freem(m);
223 		IP_STATDEC(IP_STAT_DELIVERED);
224 		/* do not inject data into pcb */
225 	} else
226 #endif
227 	if (last != NULL) {
228 		rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, m);
229 	} else if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) {
230 		uint64_t *ips;
231 
232 		icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL,
233 		    0, 0);
234 		ips = IP_STAT_GETREF();
235 		ips[IP_STAT_NOPROTO]++;
236 		ips[IP_STAT_DELIVERED]--;
237 		IP_STAT_PUTREF();
238 	} else {
239 		m_freem(m);
240 	}
241 
242 	return;
243 }
244 
245 int
246 rip_pcbnotify(struct inpcbtable *table,
247     struct in_addr faddr, struct in_addr laddr, int proto, int errno,
248     void (*notify)(struct inpcb *, int))
249 {
250 	struct inpcb_hdr *inph, *ninph;
251 	int nmatch;
252 
253 	nmatch = 0;
254 	TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) {
255 		struct inpcb *inp = (struct inpcb *)inph;
256 		if (inp->inp_af != AF_INET)
257 			continue;
258 		if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
259 			continue;
260 		if (in_hosteq(inp->inp_faddr, faddr) &&
261 		    in_hosteq(inp->inp_laddr, laddr)) {
262 			(*notify)(inp, errno);
263 			nmatch++;
264 		}
265 	}
266 
267 	return nmatch;
268 }
269 
270 void *
271 rip_ctlinput(int cmd, const struct sockaddr *sa, void *v)
272 {
273 	struct ip *ip = v;
274 	void (*notify)(struct inpcb *, int) = in_rtchange;
275 	int errno;
276 
277 	if (sa->sa_family != AF_INET ||
278 	    sa->sa_len != sizeof(struct sockaddr_in))
279 		return NULL;
280 	if ((unsigned)cmd >= PRC_NCMDS)
281 		return NULL;
282 	errno = inetctlerrmap[cmd];
283 	if (PRC_IS_REDIRECT(cmd))
284 		notify = in_rtchange, ip = 0;
285 	else if (cmd == PRC_HOSTDEAD)
286 		ip = 0;
287 	else if (errno == 0)
288 		return NULL;
289 	if (ip) {
290 		rip_pcbnotify(&rawcbtable, satocsin(sa)->sin_addr,
291 		    ip->ip_src, ip->ip_p, errno, notify);
292 
293 		/* XXX mapped address case */
294 	} else
295 		in_pcbnotifyall(&rawcbtable, satocsin(sa)->sin_addr, errno,
296 		    notify);
297 	return NULL;
298 }
299 
300 /*
301  * Generate IP header and pass packet to ip_output.
302  * Tack on options user may have setup with control call.
303  */
304 int
305 rip_output(struct mbuf *m, struct inpcb *inp, struct mbuf *control,
306     struct lwp *l)
307 {
308 	struct ip *ip;
309 	struct mbuf *opts;
310 	struct ip_pktopts pktopts;
311 	kauth_cred_t cred;
312 	int error, flags;
313 
314 	flags = (inp->inp_socket->so_options & SO_DONTROUTE) |
315 	    IP_ALLOWBROADCAST | IP_RETURNMTU;
316 
317 	if (l == NULL)
318 		cred = NULL;
319 	else
320 		cred = l->l_cred;
321 
322 	/* Setup IP outgoing packet options */
323 	memset(&pktopts, 0, sizeof(pktopts));
324 	error = ip_setpktopts(control, &pktopts, &flags, inp, cred);
325 	if (control != NULL)
326 		m_freem(control);
327 	if (error != 0)
328 		goto release;
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 			error = EMSGSIZE;
337 			goto release;
338 		}
339 		M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
340 		if (!m) {
341 			error = ENOBUFS;
342 			goto release;
343 		}
344 		ip = mtod(m, struct ip *);
345 		ip->ip_tos = 0;
346 		ip->ip_off = htons(0);
347 		ip->ip_p = inp->inp_ip.ip_p;
348 		ip->ip_len = htons(m->m_pkthdr.len);
349 		ip->ip_src = pktopts.ippo_laddr.sin_addr;
350 		ip->ip_dst = inp->inp_faddr;
351 		ip->ip_ttl = MAXTTL;
352 		opts = inp->inp_options;
353 	} else {
354 		if (m->m_pkthdr.len > IP_MAXPACKET) {
355 			error = EMSGSIZE;
356 			goto release;
357 		}
358 		ip = mtod(m, struct ip *);
359 
360 		/*
361 		 * If the mbuf is read-only, we need to allocate
362 		 * a new mbuf for the header, since we need to
363 		 * modify the header.
364 		 */
365 		if (M_READONLY(m)) {
366 			int hlen = ip->ip_hl << 2;
367 
368 			m = m_copyup(m, hlen, (max_linkhdr + 3) & ~3);
369 			if (m == NULL) {
370 				error = ENOMEM;	/* XXX */
371 				goto release;
372 			}
373 			ip = mtod(m, struct ip *);
374 		}
375 
376 		/* XXX userland passes ip_len and ip_off in host order */
377 		if (m->m_pkthdr.len != ip->ip_len) {
378 			error = EINVAL;
379 			goto release;
380 		}
381 		HTONS(ip->ip_len);
382 		HTONS(ip->ip_off);
383 		if (ip->ip_id != 0 || m->m_pkthdr.len < IP_MINFRAGSIZE)
384 			flags |= IP_NOIPNEWID;
385 		opts = NULL;
386 		/* XXX prevent ip_output from overwriting header fields */
387 		flags |= IP_RAWOUTPUT;
388 		IP_STATINC(IP_STAT_RAWOUT);
389 	}
390 
391 	/*
392 	 * IP output.  Note: if IP_RETURNMTU flag is set, the MTU size
393 	 * will be stored in inp_errormtu.
394 	 */
395 	return ip_output(m, opts, &inp->inp_route, flags, pktopts.ippo_imo,
396 	    inp);
397 
398  release:
399 	if (m != NULL)
400 		m_freem(m);
401 	return error;
402 }
403 
404 /*
405  * Raw IP socket option processing.
406  */
407 int
408 rip_ctloutput(int op, struct socket *so, struct sockopt *sopt)
409 {
410 	struct inpcb *inp = sotoinpcb(so);
411 	int error = 0;
412 	int optval;
413 
414 	if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) {
415 		if (op == PRCO_GETOPT) {
416 			optval = (inp->inp_flags & INP_NOHEADER) ? 1 : 0;
417 			error = sockopt_set(sopt, &optval, sizeof(optval));
418 		} else if (op == PRCO_SETOPT) {
419 			error = sockopt_getint(sopt, &optval);
420 			if (error)
421 				goto out;
422 			if (optval) {
423 				inp->inp_flags &= ~INP_HDRINCL;
424 				inp->inp_flags |= INP_NOHEADER;
425 			} else
426 				inp->inp_flags &= ~INP_NOHEADER;
427 		}
428 		goto out;
429 	} else if (sopt->sopt_level != IPPROTO_IP)
430 		return ip_ctloutput(op, so, sopt);
431 
432 	switch (op) {
433 
434 	case PRCO_SETOPT:
435 		switch (sopt->sopt_name) {
436 		case IP_HDRINCL:
437 			error = sockopt_getint(sopt, &optval);
438 			if (error)
439 				break;
440 			if (optval)
441 				inp->inp_flags |= INP_HDRINCL;
442 			else
443 				inp->inp_flags &= ~INP_HDRINCL;
444 			break;
445 
446 #ifdef MROUTING
447 		case MRT_INIT:
448 		case MRT_DONE:
449 		case MRT_ADD_VIF:
450 		case MRT_DEL_VIF:
451 		case MRT_ADD_MFC:
452 		case MRT_DEL_MFC:
453 		case MRT_ASSERT:
454 		case MRT_API_CONFIG:
455 		case MRT_ADD_BW_UPCALL:
456 		case MRT_DEL_BW_UPCALL:
457 			error = ip_mrouter_set(so, sopt);
458 			break;
459 #endif
460 
461 		default:
462 			error = ip_ctloutput(op, so, sopt);
463 			break;
464 		}
465 		break;
466 
467 	case PRCO_GETOPT:
468 		switch (sopt->sopt_name) {
469 		case IP_HDRINCL:
470 			optval = inp->inp_flags & INP_HDRINCL;
471 			error = sockopt_set(sopt, &optval, sizeof(optval));
472 			break;
473 
474 #ifdef MROUTING
475 		case MRT_VERSION:
476 		case MRT_ASSERT:
477 		case MRT_API_SUPPORT:
478 		case MRT_API_CONFIG:
479 			error = ip_mrouter_get(so, sopt);
480 			break;
481 #endif
482 
483 		default:
484 			error = ip_ctloutput(op, so, sopt);
485 			break;
486 		}
487 		break;
488 	}
489  out:
490 	return error;
491 }
492 
493 int
494 rip_connect_pcb(struct inpcb *inp, struct sockaddr_in *addr)
495 {
496 
497 	if (IFNET_READER_EMPTY())
498 		return (EADDRNOTAVAIL);
499 	if (addr->sin_family != AF_INET)
500 		return (EAFNOSUPPORT);
501 	inp->inp_faddr = addr->sin_addr;
502 	return (0);
503 }
504 
505 static void
506 rip_disconnect1(struct inpcb *inp)
507 {
508 
509 	inp->inp_faddr = zeroin_addr;
510 }
511 
512 static int
513 rip_attach(struct socket *so, int proto)
514 {
515 	struct inpcb *inp;
516 	int error;
517 
518 	KASSERT(sotoinpcb(so) == NULL);
519 	sosetlock(so);
520 
521 	if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
522 		error = soreserve(so, rip_sendspace, rip_recvspace);
523 		if (error) {
524 			return error;
525 		}
526 	}
527 
528 	error = in_pcballoc(so, &rawcbtable);
529 	if (error) {
530 		return error;
531 	}
532 	inp = sotoinpcb(so);
533 	inp->inp_ip.ip_p = proto;
534 	KASSERT(solocked(so));
535 
536 	return 0;
537 }
538 
539 static void
540 rip_detach(struct socket *so)
541 {
542 	struct inpcb *inp;
543 
544 	KASSERT(solocked(so));
545 	inp = sotoinpcb(so);
546 	KASSERT(inp != NULL);
547 
548 #ifdef MROUTING
549 	extern struct socket *ip_mrouter;
550 	if (so == ip_mrouter) {
551 		ip_mrouter_done();
552 	}
553 #endif
554 	in_pcbdetach(inp);
555 }
556 
557 static int
558 rip_accept(struct socket *so, struct sockaddr *nam)
559 {
560 	KASSERT(solocked(so));
561 
562 	panic("rip_accept");
563 
564 	return EOPNOTSUPP;
565 }
566 
567 static int
568 rip_bind(struct socket *so, struct sockaddr *nam, struct lwp *l)
569 {
570 	struct inpcb *inp = sotoinpcb(so);
571 	struct sockaddr_in *addr = (struct sockaddr_in *)nam;
572 	int error = 0;
573 	int s, ss;
574 	struct ifaddr *ifa;
575 
576 	KASSERT(solocked(so));
577 	KASSERT(inp != NULL);
578 	KASSERT(nam != NULL);
579 
580 	if (addr->sin_len != sizeof(*addr))
581 		return EINVAL;
582 
583 	s = splsoftnet();
584 	if (IFNET_READER_EMPTY()) {
585 		error = EADDRNOTAVAIL;
586 		goto release;
587 	}
588 	if (addr->sin_family != AF_INET) {
589 		error = EAFNOSUPPORT;
590 		goto release;
591 	}
592 	ss = pserialize_read_enter();
593 	if ((ifa = ifa_ifwithaddr(sintosa(addr))) == NULL &&
594 	    !in_nullhost(addr->sin_addr))
595 	{
596 		pserialize_read_exit(ss);
597 		error = EADDRNOTAVAIL;
598 		goto release;
599 	}
600         if (ifa && (ifatoia(ifa))->ia4_flags & IN6_IFF_DUPLICATED) {
601 		pserialize_read_exit(ss);
602 		error = EADDRNOTAVAIL;
603 		goto release;
604 	}
605 	pserialize_read_exit(ss);
606 
607 	inp->inp_laddr = addr->sin_addr;
608 
609 release:
610 	splx(s);
611 	return error;
612 }
613 
614 static int
615 rip_listen(struct socket *so, struct lwp *l)
616 {
617 	KASSERT(solocked(so));
618 
619 	return EOPNOTSUPP;
620 }
621 
622 static int
623 rip_connect(struct socket *so, struct sockaddr *nam, struct lwp *l)
624 {
625 	struct inpcb *inp = sotoinpcb(so);
626 	int error = 0;
627 	int s;
628 
629 	KASSERT(solocked(so));
630 	KASSERT(inp != NULL);
631 	KASSERT(nam != NULL);
632 
633 	s = splsoftnet();
634 	error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
635 	if (! error)
636 		soisconnected(so);
637 	splx(s);
638 
639 	return error;
640 }
641 
642 static int
643 rip_connect2(struct socket *so, struct socket *so2)
644 {
645 	KASSERT(solocked(so));
646 
647 	return EOPNOTSUPP;
648 }
649 
650 static int
651 rip_disconnect(struct socket *so)
652 {
653 	struct inpcb *inp = sotoinpcb(so);
654 	int s;
655 
656 	KASSERT(solocked(so));
657 	KASSERT(inp != NULL);
658 
659 	s = splsoftnet();
660 	soisdisconnected(so);
661 	rip_disconnect1(inp);
662 	splx(s);
663 
664 	return 0;
665 }
666 
667 static int
668 rip_shutdown(struct socket *so)
669 {
670 	int s;
671 
672 	KASSERT(solocked(so));
673 
674 	/*
675 	 * Mark the connection as being incapable of further input.
676 	 */
677 	s = splsoftnet();
678 	socantsendmore(so);
679 	splx(s);
680 
681 	return 0;
682 }
683 
684 static int
685 rip_abort(struct socket *so)
686 {
687 	KASSERT(solocked(so));
688 
689 	panic("rip_abort");
690 
691 	return EOPNOTSUPP;
692 }
693 
694 static int
695 rip_ioctl(struct socket *so, u_long cmd, void *nam, struct ifnet *ifp)
696 {
697 	return in_control(so, cmd, nam, ifp);
698 }
699 
700 static int
701 rip_stat(struct socket *so, struct stat *ub)
702 {
703 	KASSERT(solocked(so));
704 
705 	/* stat: don't bother with a blocksize. */
706 	return 0;
707 }
708 
709 static int
710 rip_peeraddr(struct socket *so, struct sockaddr *nam)
711 {
712 	int s;
713 
714 	KASSERT(solocked(so));
715 	KASSERT(sotoinpcb(so) != NULL);
716 	KASSERT(nam != NULL);
717 
718 	s = splsoftnet();
719 	in_setpeeraddr(sotoinpcb(so), (struct sockaddr_in *)nam);
720 	splx(s);
721 
722 	return 0;
723 }
724 
725 static int
726 rip_sockaddr(struct socket *so, struct sockaddr *nam)
727 {
728 	int s;
729 
730 	KASSERT(solocked(so));
731 	KASSERT(sotoinpcb(so) != NULL);
732 	KASSERT(nam != NULL);
733 
734 	s = splsoftnet();
735 	in_setsockaddr(sotoinpcb(so), (struct sockaddr_in *)nam);
736 	splx(s);
737 
738 	return 0;
739 }
740 
741 static int
742 rip_rcvd(struct socket *so, int flags, struct lwp *l)
743 {
744 	KASSERT(solocked(so));
745 
746 	return EOPNOTSUPP;
747 }
748 
749 static int
750 rip_recvoob(struct socket *so, struct mbuf *m, int flags)
751 {
752 	KASSERT(solocked(so));
753 
754 	return EOPNOTSUPP;
755 }
756 
757 static int
758 rip_send(struct socket *so, struct mbuf *m, struct sockaddr *nam,
759     struct mbuf *control, struct lwp *l)
760 {
761 	struct inpcb *inp = sotoinpcb(so);
762 	int error = 0;
763 	int s;
764 
765 	KASSERT(solocked(so));
766 	KASSERT(inp != NULL);
767 	KASSERT(m != NULL);
768 
769 	/*
770 	 * Ship a packet out.  The appropriate raw output
771 	 * routine handles any massaging necessary.
772 	 */
773 	s = splsoftnet();
774 	if (nam) {
775 		if ((so->so_state & SS_ISCONNECTED) != 0) {
776 			error = EISCONN;
777 			goto die;
778 		}
779 		error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
780 		if (error)
781 			goto die;
782 	} else {
783 		if ((so->so_state & SS_ISCONNECTED) == 0) {
784 			error = ENOTCONN;
785 			goto die;
786 		}
787 	}
788 	error = rip_output(m, inp, control, l);
789 	m = NULL;
790 	control = NULL;
791 	if (nam)
792 		rip_disconnect1(inp);
793  die:
794 	if (m != NULL)
795 		m_freem(m);
796 	if (control != NULL)
797 		m_freem(control);
798 
799 	splx(s);
800 	return error;
801 }
802 
803 static int
804 rip_sendoob(struct socket *so, struct mbuf *m, struct mbuf *control)
805 {
806 	KASSERT(solocked(so));
807 
808 	m_freem(m);
809 	m_freem(control);
810 
811 	return EOPNOTSUPP;
812 }
813 
814 static int
815 rip_purgeif(struct socket *so, struct ifnet *ifp)
816 {
817 	int s;
818 
819 	s = splsoftnet();
820 	mutex_enter(softnet_lock);
821 	in_pcbpurgeif0(&rawcbtable, ifp);
822 #ifdef NET_MPSAFE
823 	mutex_exit(softnet_lock);
824 #endif
825 	in_purgeif(ifp);
826 #ifdef NET_MPSAFE
827 	mutex_enter(softnet_lock);
828 #endif
829 	in_pcbpurgeif(&rawcbtable, ifp);
830 	mutex_exit(softnet_lock);
831 	splx(s);
832 
833 	return 0;
834 }
835 
836 PR_WRAP_USRREQS(rip)
837 #define	rip_attach	rip_attach_wrapper
838 #define	rip_detach	rip_detach_wrapper
839 #define	rip_accept	rip_accept_wrapper
840 #define	rip_bind	rip_bind_wrapper
841 #define	rip_listen	rip_listen_wrapper
842 #define	rip_connect	rip_connect_wrapper
843 #define	rip_connect2	rip_connect2_wrapper
844 #define	rip_disconnect	rip_disconnect_wrapper
845 #define	rip_shutdown	rip_shutdown_wrapper
846 #define	rip_abort	rip_abort_wrapper
847 #define	rip_ioctl	rip_ioctl_wrapper
848 #define	rip_stat	rip_stat_wrapper
849 #define	rip_peeraddr	rip_peeraddr_wrapper
850 #define	rip_sockaddr	rip_sockaddr_wrapper
851 #define	rip_rcvd	rip_rcvd_wrapper
852 #define	rip_recvoob	rip_recvoob_wrapper
853 #define	rip_send	rip_send_wrapper
854 #define	rip_sendoob	rip_sendoob_wrapper
855 #define	rip_purgeif	rip_purgeif_wrapper
856 
857 const struct pr_usrreqs rip_usrreqs = {
858 	.pr_attach	= rip_attach,
859 	.pr_detach	= rip_detach,
860 	.pr_accept	= rip_accept,
861 	.pr_bind	= rip_bind,
862 	.pr_listen	= rip_listen,
863 	.pr_connect	= rip_connect,
864 	.pr_connect2	= rip_connect2,
865 	.pr_disconnect	= rip_disconnect,
866 	.pr_shutdown	= rip_shutdown,
867 	.pr_abort	= rip_abort,
868 	.pr_ioctl	= rip_ioctl,
869 	.pr_stat	= rip_stat,
870 	.pr_peeraddr	= rip_peeraddr,
871 	.pr_sockaddr	= rip_sockaddr,
872 	.pr_rcvd	= rip_rcvd,
873 	.pr_recvoob	= rip_recvoob,
874 	.pr_send	= rip_send,
875 	.pr_sendoob	= rip_sendoob,
876 	.pr_purgeif	= rip_purgeif,
877 };
878 
879 static void
880 sysctl_net_inet_raw_setup(struct sysctllog **clog)
881 {
882 
883 	sysctl_createv(clog, 0, NULL, NULL,
884 		       CTLFLAG_PERMANENT,
885 		       CTLTYPE_NODE, "inet", NULL,
886 		       NULL, 0, NULL, 0,
887 		       CTL_NET, PF_INET, CTL_EOL);
888 	sysctl_createv(clog, 0, NULL, NULL,
889 		       CTLFLAG_PERMANENT,
890 		       CTLTYPE_NODE, "raw",
891 		       SYSCTL_DESCR("Raw IPv4 settings"),
892 		       NULL, 0, NULL, 0,
893 		       CTL_NET, PF_INET, IPPROTO_RAW, CTL_EOL);
894 
895 	sysctl_createv(clog, 0, NULL, NULL,
896 		       CTLFLAG_PERMANENT,
897 		       CTLTYPE_STRUCT, "pcblist",
898 		       SYSCTL_DESCR("Raw IPv4 control block list"),
899 		       sysctl_inpcblist, 0, &rawcbtable, 0,
900 		       CTL_NET, PF_INET, IPPROTO_RAW,
901 		       CTL_CREATE, CTL_EOL);
902 }
903