xref: /netbsd-src/sys/netinet/raw_ip.c (revision f4748aaa01faf324805f9747191535eb6600f82c)
1 /*	$NetBSD: raw_ip.c,v 1.184 2022/11/04 09:00:58 ozaki-r 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 1.184 2022/11/04 09:00:58 ozaki-r 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 	inpcb_init(&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, int off, int proto)
166 {
167 	struct ip *ip = mtod(m, struct ip *);
168 	struct inpcb *inp;
169 	struct inpcb *last = NULL;
170 	struct mbuf *n;
171 	struct sockaddr_in ripsrc;
172 	int hlen;
173 
174 	sockaddr_in_init(&ripsrc, &ip->ip_src, 0);
175 
176 	/*
177 	 * XXX Compatibility: programs using raw IP expect ip_len
178 	 * XXX to have the header length subtracted, and in host order.
179 	 * XXX ip_off is also expected to be host order.
180 	 */
181 	hlen = ip->ip_hl << 2;
182 	ip->ip_len = ntohs(ip->ip_len) - hlen;
183 	NTOHS(ip->ip_off);
184 
185 	TAILQ_FOREACH(inp, &rawcbtable.inpt_queue, inp_queue) {
186 		if (inp->inp_af != AF_INET)
187 			continue;
188 		if (in4p_ip(inp).ip_p && in4p_ip(inp).ip_p != proto)
189 			continue;
190 		if (!in_nullhost(in4p_laddr(inp)) &&
191 		    !in_hosteq(in4p_laddr(inp), ip->ip_dst))
192 			continue;
193 		if (!in_nullhost(in4p_faddr(inp)) &&
194 		    !in_hosteq(in4p_faddr(inp), ip->ip_src))
195 			continue;
196 
197 		if (last == NULL) {
198 			;
199 		}
200 #if defined(IPSEC)
201 		else if (ipsec_used && ipsec_in_reject(m, last)) {
202 			/* do not inject data into pcb */
203 		}
204 #endif
205 		else if ((n = m_copypacket(m, M_DONTWAIT)) != NULL) {
206 			rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, n);
207 		}
208 
209 		last = inp;
210 	}
211 
212 #if defined(IPSEC)
213 	if (ipsec_used && last != NULL && ipsec_in_reject(m, last)) {
214 		m_freem(m);
215 		IP_STATDEC(IP_STAT_DELIVERED);
216 		/* do not inject data into pcb */
217 	} else
218 #endif
219 	if (last != NULL) {
220 		rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, m);
221 	} else if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) {
222 		uint64_t *ips;
223 
224 		icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL,
225 		    0, 0);
226 		ips = IP_STAT_GETREF();
227 		ips[IP_STAT_NOPROTO]++;
228 		ips[IP_STAT_DELIVERED]--;
229 		IP_STAT_PUTREF();
230 	} else {
231 		m_freem(m);
232 	}
233 
234 	return;
235 }
236 
237 int
238 rip_pcbnotify(struct inpcbtable *table,
239     struct in_addr faddr, struct in_addr laddr, int proto, int errno,
240     void (*notify)(struct inpcb *, int))
241 {
242 	struct inpcb *inp;
243 	int nmatch;
244 
245 	nmatch = 0;
246 	TAILQ_FOREACH(inp, &table->inpt_queue, inp_queue) {
247 		if (inp->inp_af != AF_INET)
248 			continue;
249 		if (in4p_ip(inp).ip_p && in4p_ip(inp).ip_p != proto)
250 			continue;
251 		if (in_hosteq(in4p_faddr(inp), faddr) &&
252 		    in_hosteq(in4p_laddr(inp), laddr)) {
253 			(*notify)(inp, errno);
254 			nmatch++;
255 		}
256 	}
257 
258 	return nmatch;
259 }
260 
261 void *
262 rip_ctlinput(int cmd, const struct sockaddr *sa, void *v)
263 {
264 	struct ip *ip = v;
265 	void (*notify)(struct inpcb *, int) = inpcb_rtchange;
266 	int errno;
267 
268 	if (sa->sa_family != AF_INET ||
269 	    sa->sa_len != sizeof(struct sockaddr_in))
270 		return NULL;
271 	if ((unsigned)cmd >= PRC_NCMDS)
272 		return NULL;
273 	errno = inetctlerrmap[cmd];
274 	if (PRC_IS_REDIRECT(cmd))
275 		notify = inpcb_rtchange, ip = 0;
276 	else if (cmd == PRC_HOSTDEAD)
277 		ip = 0;
278 	else if (errno == 0)
279 		return NULL;
280 	if (ip) {
281 		rip_pcbnotify(&rawcbtable, satocsin(sa)->sin_addr,
282 		    ip->ip_src, ip->ip_p, errno, notify);
283 
284 		/* XXX mapped address case */
285 	} else
286 		inpcb_notifyall(&rawcbtable, satocsin(sa)->sin_addr, errno,
287 		    notify);
288 	return NULL;
289 }
290 
291 /*
292  * Generate IP header and pass packet to ip_output.
293  * Tack on options user may have setup with control call.
294  */
295 int
296 rip_output(struct mbuf *m, struct inpcb *inp, struct mbuf *control,
297     struct lwp *l)
298 {
299 	struct ip *ip;
300 	struct mbuf *opts;
301 	struct ip_pktopts pktopts;
302 	kauth_cred_t cred;
303 	int error, flags;
304 
305 	flags = (inp->inp_socket->so_options & SO_DONTROUTE) |
306 	    IP_ALLOWBROADCAST | IP_RETURNMTU;
307 
308 	if (l == NULL)
309 		cred = NULL;
310 	else
311 		cred = l->l_cred;
312 
313 	/* Setup IP outgoing packet options */
314 	memset(&pktopts, 0, sizeof(pktopts));
315 	error = ip_setpktopts(control, &pktopts, &flags, inp, cred);
316 	if (control != NULL)
317 		m_freem(control);
318 	if (error != 0)
319 		goto release;
320 
321 	/*
322 	 * If the user handed us a complete IP packet, use it.
323 	 * Otherwise, allocate an mbuf for a header and fill it in.
324 	 */
325 	if ((inp->inp_flags & INP_HDRINCL) == 0) {
326 		if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
327 			error = EMSGSIZE;
328 			goto release;
329 		}
330 		M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
331 		if (!m) {
332 			error = ENOBUFS;
333 			goto release;
334 		}
335 		ip = mtod(m, struct ip *);
336 		ip->ip_tos = 0;
337 		ip->ip_off = htons(0);
338 		ip->ip_p = in4p_ip(inp).ip_p;
339 		ip->ip_len = htons(m->m_pkthdr.len);
340 		ip->ip_src = pktopts.ippo_laddr.sin_addr;
341 		ip->ip_dst = in4p_faddr(inp);
342 		ip->ip_ttl = MAXTTL;
343 		opts = inp->inp_options;
344 	} else {
345 		if (m->m_pkthdr.len > IP_MAXPACKET) {
346 			error = EMSGSIZE;
347 			goto release;
348 		}
349 		if (m->m_pkthdr.len < sizeof(struct ip)) {
350 			error = EINVAL;
351 			goto release;
352 		}
353 		ip = mtod(m, struct ip *);
354 
355 		/*
356 		 * If the mbuf is read-only, we need to allocate
357 		 * a new mbuf for the header, since we need to
358 		 * modify the header.
359 		 */
360 		if (M_READONLY(m)) {
361 			int hlen = ip->ip_hl << 2;
362 
363 			m = m_copyup(m, hlen, (max_linkhdr + 3) & ~3);
364 			if (m == NULL) {
365 				error = ENOMEM;
366 				goto release;
367 			}
368 			ip = mtod(m, struct ip *);
369 		}
370 
371 		/* XXX userland passes ip_len and ip_off in host order */
372 		if (m->m_pkthdr.len != ip->ip_len) {
373 			error = EINVAL;
374 			goto release;
375 		}
376 		HTONS(ip->ip_len);
377 		HTONS(ip->ip_off);
378 
379 		if (ip->ip_id != 0 || m->m_pkthdr.len < IP_MINFRAGSIZE)
380 			flags |= IP_NOIPNEWID;
381 		opts = NULL;
382 
383 		/* Prevent ip_output from overwriting header fields. */
384 		flags |= IP_RAWOUTPUT;
385 
386 		IP_STATINC(IP_STAT_RAWOUT);
387 	}
388 
389 	/*
390 	 * IP output.  Note: if IP_RETURNMTU flag is set, the MTU size
391 	 * will be stored in inp_errormtu.
392 	 */
393 	return ip_output(m, opts, &inp->inp_route, flags, pktopts.ippo_imo,
394 	    inp);
395 
396  release:
397 	if (m != NULL)
398 		m_freem(m);
399 	return error;
400 }
401 
402 /*
403  * Raw IP socket option processing.
404  */
405 int
406 rip_ctloutput(int op, struct socket *so, struct sockopt *sopt)
407 {
408 	struct inpcb *inp = sotoinpcb(so);
409 	int error = 0;
410 	int optval;
411 
412 	if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) {
413 		if (op == PRCO_GETOPT) {
414 			optval = (inp->inp_flags & INP_NOHEADER) ? 1 : 0;
415 			error = sockopt_set(sopt, &optval, sizeof(optval));
416 		} else if (op == PRCO_SETOPT) {
417 			error = sockopt_getint(sopt, &optval);
418 			if (error)
419 				goto out;
420 			if (optval) {
421 				inp->inp_flags &= ~INP_HDRINCL;
422 				inp->inp_flags |= INP_NOHEADER;
423 			} else
424 				inp->inp_flags &= ~INP_NOHEADER;
425 		}
426 		goto out;
427 	} else if (sopt->sopt_level != IPPROTO_IP)
428 		return ip_ctloutput(op, so, sopt);
429 
430 	switch (op) {
431 
432 	case PRCO_SETOPT:
433 		switch (sopt->sopt_name) {
434 		case IP_HDRINCL:
435 			error = sockopt_getint(sopt, &optval);
436 			if (error)
437 				break;
438 			if (optval)
439 				inp->inp_flags |= INP_HDRINCL;
440 			else
441 				inp->inp_flags &= ~INP_HDRINCL;
442 			break;
443 
444 #ifdef MROUTING
445 		case MRT_INIT:
446 		case MRT_DONE:
447 		case MRT_ADD_VIF:
448 		case MRT_DEL_VIF:
449 		case MRT_ADD_MFC:
450 		case MRT_DEL_MFC:
451 		case MRT_ASSERT:
452 		case MRT_API_CONFIG:
453 		case MRT_ADD_BW_UPCALL:
454 		case MRT_DEL_BW_UPCALL:
455 			error = ip_mrouter_set(so, sopt);
456 			break;
457 #endif
458 
459 		default:
460 			error = ip_ctloutput(op, so, sopt);
461 			break;
462 		}
463 		break;
464 
465 	case PRCO_GETOPT:
466 		switch (sopt->sopt_name) {
467 		case IP_HDRINCL:
468 			optval = inp->inp_flags & INP_HDRINCL;
469 			error = sockopt_set(sopt, &optval, sizeof(optval));
470 			break;
471 
472 #ifdef MROUTING
473 		case MRT_VERSION:
474 		case MRT_ASSERT:
475 		case MRT_API_SUPPORT:
476 		case MRT_API_CONFIG:
477 			error = ip_mrouter_get(so, sopt);
478 			break;
479 #endif
480 
481 		default:
482 			error = ip_ctloutput(op, so, sopt);
483 			break;
484 		}
485 		break;
486 	}
487  out:
488 	return error;
489 }
490 
491 int
492 rip_connect_pcb(struct inpcb *inp, struct sockaddr_in *addr)
493 {
494 
495 	if (IFNET_READER_EMPTY())
496 		return (EADDRNOTAVAIL);
497 	if (addr->sin_family != AF_INET)
498 		return (EAFNOSUPPORT);
499 	if (addr->sin_len != sizeof(*addr))
500 		return EINVAL;
501 	in4p_faddr(inp) = addr->sin_addr;
502 	return (0);
503 }
504 
505 static void
506 rip_disconnect1(struct inpcb *inp)
507 {
508 
509 	in4p_faddr(inp) = 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 = inpcb_create(so, &rawcbtable);
529 	if (error) {
530 		return error;
531 	}
532 	inp = sotoinpcb(so);
533 	in4p_ip(inp).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 	inpcb_destroy(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 	    (inp->inp_flags & INP_BINDANY) == 0 &&
595 	    !in_nullhost(addr->sin_addr))
596 	{
597 		pserialize_read_exit(ss);
598 		error = EADDRNOTAVAIL;
599 		goto release;
600 	}
601         if (ifa && (ifatoia(ifa))->ia4_flags & IN6_IFF_DUPLICATED) {
602 		pserialize_read_exit(ss);
603 		error = EADDRNOTAVAIL;
604 		goto release;
605 	}
606 	pserialize_read_exit(ss);
607 
608 	in4p_laddr(inp) = addr->sin_addr;
609 
610 release:
611 	splx(s);
612 	return error;
613 }
614 
615 static int
616 rip_listen(struct socket *so, struct lwp *l)
617 {
618 	KASSERT(solocked(so));
619 
620 	return EOPNOTSUPP;
621 }
622 
623 static int
624 rip_connect(struct socket *so, struct sockaddr *nam, struct lwp *l)
625 {
626 	struct inpcb *inp = sotoinpcb(so);
627 	int error = 0;
628 	int s;
629 
630 	KASSERT(solocked(so));
631 	KASSERT(inp != NULL);
632 	KASSERT(nam != NULL);
633 
634 	s = splsoftnet();
635 	error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
636 	if (! error)
637 		soisconnected(so);
638 	splx(s);
639 
640 	return error;
641 }
642 
643 static int
644 rip_connect2(struct socket *so, struct socket *so2)
645 {
646 	KASSERT(solocked(so));
647 
648 	return EOPNOTSUPP;
649 }
650 
651 static int
652 rip_disconnect(struct socket *so)
653 {
654 	struct inpcb *inp = sotoinpcb(so);
655 	int s;
656 
657 	KASSERT(solocked(so));
658 	KASSERT(inp != NULL);
659 
660 	s = splsoftnet();
661 	soisdisconnected(so);
662 	rip_disconnect1(inp);
663 	splx(s);
664 
665 	return 0;
666 }
667 
668 static int
669 rip_shutdown(struct socket *so)
670 {
671 	int s;
672 
673 	KASSERT(solocked(so));
674 
675 	/*
676 	 * Mark the connection as being incapable of further input.
677 	 */
678 	s = splsoftnet();
679 	socantsendmore(so);
680 	splx(s);
681 
682 	return 0;
683 }
684 
685 static int
686 rip_abort(struct socket *so)
687 {
688 	KASSERT(solocked(so));
689 
690 	panic("rip_abort");
691 
692 	return EOPNOTSUPP;
693 }
694 
695 static int
696 rip_ioctl(struct socket *so, u_long cmd, void *nam, struct ifnet *ifp)
697 {
698 	return in_control(so, cmd, nam, ifp);
699 }
700 
701 static int
702 rip_stat(struct socket *so, struct stat *ub)
703 {
704 	KASSERT(solocked(so));
705 
706 	/* stat: don't bother with a blocksize. */
707 	return 0;
708 }
709 
710 static int
711 rip_peeraddr(struct socket *so, struct sockaddr *nam)
712 {
713 	int s;
714 
715 	KASSERT(solocked(so));
716 	KASSERT(sotoinpcb(so) != NULL);
717 	KASSERT(nam != NULL);
718 
719 	s = splsoftnet();
720 	inpcb_fetch_peeraddr(sotoinpcb(so), (struct sockaddr_in *)nam);
721 	splx(s);
722 
723 	return 0;
724 }
725 
726 static int
727 rip_sockaddr(struct socket *so, struct sockaddr *nam)
728 {
729 	int s;
730 
731 	KASSERT(solocked(so));
732 	KASSERT(sotoinpcb(so) != NULL);
733 	KASSERT(nam != NULL);
734 
735 	s = splsoftnet();
736 	inpcb_fetch_sockaddr(sotoinpcb(so), (struct sockaddr_in *)nam);
737 	splx(s);
738 
739 	return 0;
740 }
741 
742 static int
743 rip_rcvd(struct socket *so, int flags, struct lwp *l)
744 {
745 	KASSERT(solocked(so));
746 
747 	return EOPNOTSUPP;
748 }
749 
750 static int
751 rip_recvoob(struct socket *so, struct mbuf *m, int flags)
752 {
753 	KASSERT(solocked(so));
754 
755 	return EOPNOTSUPP;
756 }
757 
758 static int
759 rip_send(struct socket *so, struct mbuf *m, struct sockaddr *nam,
760     struct mbuf *control, struct lwp *l)
761 {
762 	struct inpcb *inp = sotoinpcb(so);
763 	int error = 0;
764 	int s;
765 
766 	KASSERT(solocked(so));
767 	KASSERT(inp != NULL);
768 	KASSERT(m != NULL);
769 
770 	/*
771 	 * Ship a packet out.  The appropriate raw output
772 	 * routine handles any massaging necessary.
773 	 */
774 	s = splsoftnet();
775 	if (nam) {
776 		if ((so->so_state & SS_ISCONNECTED) != 0) {
777 			error = EISCONN;
778 			goto die;
779 		}
780 		error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
781 		if (error)
782 			goto die;
783 	} else {
784 		if ((so->so_state & SS_ISCONNECTED) == 0) {
785 			error = ENOTCONN;
786 			goto die;
787 		}
788 	}
789 	error = rip_output(m, inp, control, l);
790 	m = NULL;
791 	control = NULL;
792 	if (nam)
793 		rip_disconnect1(inp);
794  die:
795 	if (m != NULL)
796 		m_freem(m);
797 	if (control != NULL)
798 		m_freem(control);
799 
800 	splx(s);
801 	return error;
802 }
803 
804 static int
805 rip_sendoob(struct socket *so, struct mbuf *m, struct mbuf *control)
806 {
807 	KASSERT(solocked(so));
808 
809 	m_freem(m);
810 	m_freem(control);
811 
812 	return EOPNOTSUPP;
813 }
814 
815 static int
816 rip_purgeif(struct socket *so, struct ifnet *ifp)
817 {
818 	int s;
819 
820 	s = splsoftnet();
821 	mutex_enter(softnet_lock);
822 	inpcb_purgeif0(&rawcbtable, ifp);
823 #ifdef NET_MPSAFE
824 	mutex_exit(softnet_lock);
825 #endif
826 	in_purgeif(ifp);
827 #ifdef NET_MPSAFE
828 	mutex_enter(softnet_lock);
829 #endif
830 	inpcb_purgeif(&rawcbtable, ifp);
831 	mutex_exit(softnet_lock);
832 	splx(s);
833 
834 	return 0;
835 }
836 
837 PR_WRAP_USRREQS(rip)
838 #define	rip_attach	rip_attach_wrapper
839 #define	rip_detach	rip_detach_wrapper
840 #define	rip_accept	rip_accept_wrapper
841 #define	rip_bind	rip_bind_wrapper
842 #define	rip_listen	rip_listen_wrapper
843 #define	rip_connect	rip_connect_wrapper
844 #define	rip_connect2	rip_connect2_wrapper
845 #define	rip_disconnect	rip_disconnect_wrapper
846 #define	rip_shutdown	rip_shutdown_wrapper
847 #define	rip_abort	rip_abort_wrapper
848 #define	rip_ioctl	rip_ioctl_wrapper
849 #define	rip_stat	rip_stat_wrapper
850 #define	rip_peeraddr	rip_peeraddr_wrapper
851 #define	rip_sockaddr	rip_sockaddr_wrapper
852 #define	rip_rcvd	rip_rcvd_wrapper
853 #define	rip_recvoob	rip_recvoob_wrapper
854 #define	rip_send	rip_send_wrapper
855 #define	rip_sendoob	rip_sendoob_wrapper
856 #define	rip_purgeif	rip_purgeif_wrapper
857 
858 const struct pr_usrreqs rip_usrreqs = {
859 	.pr_attach	= rip_attach,
860 	.pr_detach	= rip_detach,
861 	.pr_accept	= rip_accept,
862 	.pr_bind	= rip_bind,
863 	.pr_listen	= rip_listen,
864 	.pr_connect	= rip_connect,
865 	.pr_connect2	= rip_connect2,
866 	.pr_disconnect	= rip_disconnect,
867 	.pr_shutdown	= rip_shutdown,
868 	.pr_abort	= rip_abort,
869 	.pr_ioctl	= rip_ioctl,
870 	.pr_stat	= rip_stat,
871 	.pr_peeraddr	= rip_peeraddr,
872 	.pr_sockaddr	= rip_sockaddr,
873 	.pr_rcvd	= rip_rcvd,
874 	.pr_recvoob	= rip_recvoob,
875 	.pr_send	= rip_send,
876 	.pr_sendoob	= rip_sendoob,
877 	.pr_purgeif	= rip_purgeif,
878 };
879 
880 static void
881 sysctl_net_inet_raw_setup(struct sysctllog **clog)
882 {
883 
884 	sysctl_createv(clog, 0, NULL, NULL,
885 		       CTLFLAG_PERMANENT,
886 		       CTLTYPE_NODE, "inet", NULL,
887 		       NULL, 0, NULL, 0,
888 		       CTL_NET, PF_INET, CTL_EOL);
889 	sysctl_createv(clog, 0, NULL, NULL,
890 		       CTLFLAG_PERMANENT,
891 		       CTLTYPE_NODE, "raw",
892 		       SYSCTL_DESCR("Raw IPv4 settings"),
893 		       NULL, 0, NULL, 0,
894 		       CTL_NET, PF_INET, IPPROTO_RAW, CTL_EOL);
895 
896 	sysctl_createv(clog, 0, NULL, NULL,
897 		       CTLFLAG_PERMANENT,
898 		       CTLTYPE_STRUCT, "pcblist",
899 		       SYSCTL_DESCR("Raw IPv4 control block list"),
900 		       sysctl_inpcblist, 0, &rawcbtable, 0,
901 		       CTL_NET, PF_INET, IPPROTO_RAW,
902 		       CTL_CREATE, CTL_EOL);
903 }
904