raw_ip.c revision 1.145.2.1 1 1.145.2.1 martin /* $NetBSD: raw_ip.c,v 1.145.2.1 2015/01/17 12:10:53 martin Exp $ */
2 1.43 itojun
3 1.43 itojun /*
4 1.43 itojun * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5 1.43 itojun * All rights reserved.
6 1.61 itojun *
7 1.43 itojun * Redistribution and use in source and binary forms, with or without
8 1.43 itojun * modification, are permitted provided that the following conditions
9 1.43 itojun * are met:
10 1.43 itojun * 1. Redistributions of source code must retain the above copyright
11 1.43 itojun * notice, this list of conditions and the following disclaimer.
12 1.43 itojun * 2. Redistributions in binary form must reproduce the above copyright
13 1.43 itojun * notice, this list of conditions and the following disclaimer in the
14 1.43 itojun * documentation and/or other materials provided with the distribution.
15 1.43 itojun * 3. Neither the name of the project nor the names of its contributors
16 1.43 itojun * may be used to endorse or promote products derived from this software
17 1.43 itojun * without specific prior written permission.
18 1.61 itojun *
19 1.43 itojun * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
20 1.43 itojun * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 1.43 itojun * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 1.43 itojun * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
23 1.43 itojun * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 1.43 itojun * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 1.43 itojun * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 1.43 itojun * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 1.43 itojun * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 1.43 itojun * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 1.43 itojun * SUCH DAMAGE.
30 1.43 itojun */
31 1.14 cgd
32 1.1 cgd /*
33 1.13 mycroft * Copyright (c) 1982, 1986, 1988, 1993
34 1.13 mycroft * The Regents of the University of California. All rights reserved.
35 1.1 cgd *
36 1.1 cgd * Redistribution and use in source and binary forms, with or without
37 1.1 cgd * modification, are permitted provided that the following conditions
38 1.1 cgd * are met:
39 1.1 cgd * 1. Redistributions of source code must retain the above copyright
40 1.1 cgd * notice, this list of conditions and the following disclaimer.
41 1.1 cgd * 2. Redistributions in binary form must reproduce the above copyright
42 1.1 cgd * notice, this list of conditions and the following disclaimer in the
43 1.1 cgd * documentation and/or other materials provided with the distribution.
44 1.71 agc * 3. Neither the name of the University nor the names of its contributors
45 1.1 cgd * may be used to endorse or promote products derived from this software
46 1.1 cgd * without specific prior written permission.
47 1.1 cgd *
48 1.1 cgd * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
49 1.1 cgd * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
50 1.1 cgd * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
51 1.1 cgd * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
52 1.1 cgd * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
53 1.1 cgd * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
54 1.1 cgd * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
55 1.1 cgd * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
56 1.1 cgd * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
57 1.1 cgd * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58 1.1 cgd * SUCH DAMAGE.
59 1.1 cgd *
60 1.39 thorpej * @(#)raw_ip.c 8.7 (Berkeley) 5/15/95
61 1.1 cgd */
62 1.59 lukem
63 1.121 rmind /*
64 1.121 rmind * Raw interface to IP protocol.
65 1.121 rmind */
66 1.121 rmind
67 1.59 lukem #include <sys/cdefs.h>
68 1.145.2.1 martin __KERNEL_RCSID(0, "$NetBSD: raw_ip.c,v 1.145.2.1 2015/01/17 12:10:53 martin Exp $");
69 1.40 scottr
70 1.78 jonathan #include "opt_inet.h"
71 1.109 christos #include "opt_compat_netbsd.h"
72 1.45 thorpej #include "opt_ipsec.h"
73 1.40 scottr #include "opt_mrouting.h"
74 1.1 cgd
75 1.7 mycroft #include <sys/param.h>
76 1.84 atatat #include <sys/sysctl.h>
77 1.7 mycroft #include <sys/mbuf.h>
78 1.7 mycroft #include <sys/socket.h>
79 1.7 mycroft #include <sys/protosw.h>
80 1.7 mycroft #include <sys/socketvar.h>
81 1.7 mycroft #include <sys/errno.h>
82 1.13 mycroft #include <sys/systm.h>
83 1.26 mycroft #include <sys/proc.h>
84 1.89 elad #include <sys/kauth.h>
85 1.1 cgd
86 1.7 mycroft #include <net/if.h>
87 1.7 mycroft #include <net/route.h>
88 1.1 cgd
89 1.7 mycroft #include <netinet/in.h>
90 1.7 mycroft #include <netinet/in_systm.h>
91 1.7 mycroft #include <netinet/ip.h>
92 1.7 mycroft #include <netinet/ip_var.h>
93 1.105 thorpej #include <netinet/ip_private.h>
94 1.13 mycroft #include <netinet/ip_mroute.h>
95 1.44 darrenr #include <netinet/ip_icmp.h>
96 1.7 mycroft #include <netinet/in_pcb.h>
97 1.87 yamt #include <netinet/in_proto.h>
98 1.24 christos #include <netinet/in_var.h>
99 1.24 christos
100 1.116 christos #ifdef IPSEC
101 1.72 jonathan #include <netipsec/ipsec.h>
102 1.106 thorpej #include <netipsec/ipsec_var.h>
103 1.106 thorpej #include <netipsec/ipsec_private.h>
104 1.116 christos #endif /* IPSEC */
105 1.72 jonathan
106 1.109 christos #ifdef COMPAT_50
107 1.109 christos #include <compat/sys/socket.h>
108 1.109 christos #endif
109 1.109 christos
110 1.20 mycroft struct inpcbtable rawcbtable;
111 1.30 pk
112 1.82 perry int rip_pcbnotify(struct inpcbtable *, struct in_addr,
113 1.82 perry struct in_addr, int, int, void (*)(struct inpcb *, int));
114 1.137 rtr int rip_connect_pcb(struct inpcb *, struct mbuf *);
115 1.139 rtr static void rip_disconnect1(struct inpcb *);
116 1.13 mycroft
117 1.110 pooka static void sysctl_net_inet_raw_setup(struct sysctllog **);
118 1.110 pooka
119 1.13 mycroft /*
120 1.13 mycroft * Nominal space allocated to a raw ip socket.
121 1.13 mycroft */
122 1.13 mycroft #define RIPSNDQ 8192
123 1.13 mycroft #define RIPRCVQ 8192
124 1.1 cgd
125 1.121 rmind static u_long rip_sendspace = RIPSNDQ;
126 1.121 rmind static u_long rip_recvspace = RIPRCVQ;
127 1.121 rmind
128 1.1 cgd /*
129 1.1 cgd * Raw interface to IP protocol.
130 1.1 cgd */
131 1.13 mycroft
132 1.13 mycroft /*
133 1.13 mycroft * Initialize raw connection block q.
134 1.13 mycroft */
135 1.13 mycroft void
136 1.83 perry rip_init(void)
137 1.13 mycroft {
138 1.13 mycroft
139 1.110 pooka sysctl_net_inet_raw_setup(NULL);
140 1.33 mycroft in_pcbinit(&rawcbtable, 1, 1);
141 1.13 mycroft }
142 1.13 mycroft
143 1.100 dyoung static void
144 1.100 dyoung rip_sbappendaddr(struct inpcb *last, struct ip *ip, const struct sockaddr *sa,
145 1.100 dyoung int hlen, struct mbuf *opts, struct mbuf *n)
146 1.100 dyoung {
147 1.100 dyoung if (last->inp_flags & INP_NOHEADER)
148 1.100 dyoung m_adj(n, hlen);
149 1.109 christos if (last->inp_flags & INP_CONTROLOPTS
150 1.109 christos #ifdef SO_OTIMESTAMP
151 1.109 christos || last->inp_socket->so_options & SO_OTIMESTAMP
152 1.109 christos #endif
153 1.109 christos || last->inp_socket->so_options & SO_TIMESTAMP)
154 1.100 dyoung ip_savecontrol(last, &opts, ip, n);
155 1.100 dyoung if (sbappendaddr(&last->inp_socket->so_rcv, sa, n, opts) == 0) {
156 1.100 dyoung /* should notify about lost packet */
157 1.100 dyoung m_freem(n);
158 1.100 dyoung if (opts)
159 1.100 dyoung m_freem(opts);
160 1.100 dyoung } else
161 1.100 dyoung sorwakeup(last->inp_socket);
162 1.100 dyoung }
163 1.100 dyoung
164 1.1 cgd /*
165 1.1 cgd * Setup generic address and protocol structures
166 1.1 cgd * for raw_input routine, then pass them along with
167 1.1 cgd * mbuf chain.
168 1.1 cgd */
169 1.9 mycroft void
170 1.24 christos rip_input(struct mbuf *m, ...)
171 1.1 cgd {
172 1.100 dyoung int hlen, proto;
173 1.53 augustss struct ip *ip = mtod(m, struct ip *);
174 1.75 itojun struct inpcb_hdr *inph;
175 1.53 augustss struct inpcb *inp;
176 1.97 dyoung struct inpcb *last = NULL;
177 1.97 dyoung struct mbuf *n, *opts = NULL;
178 1.32 mycroft struct sockaddr_in ripsrc;
179 1.43 itojun va_list ap;
180 1.43 itojun
181 1.43 itojun va_start(ap, m);
182 1.64 simonb (void)va_arg(ap, int); /* ignore value, advance ap */
183 1.43 itojun proto = va_arg(ap, int);
184 1.43 itojun va_end(ap);
185 1.1 cgd
186 1.97 dyoung sockaddr_in_init(&ripsrc, &ip->ip_src, 0);
187 1.42 thorpej
188 1.42 thorpej /*
189 1.42 thorpej * XXX Compatibility: programs using raw IP expect ip_len
190 1.62 itojun * XXX to have the header length subtracted, and in host order.
191 1.62 itojun * XXX ip_off is also expected to be host order.
192 1.42 thorpej */
193 1.100 dyoung hlen = ip->ip_hl << 2;
194 1.100 dyoung ip->ip_len = ntohs(ip->ip_len) - hlen;
195 1.62 itojun NTOHS(ip->ip_off);
196 1.32 mycroft
197 1.117 christos TAILQ_FOREACH(inph, &rawcbtable.inpt_queue, inph_queue) {
198 1.75 itojun inp = (struct inpcb *)inph;
199 1.75 itojun if (inp->inp_af != AF_INET)
200 1.75 itojun continue;
201 1.43 itojun if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
202 1.13 mycroft continue;
203 1.32 mycroft if (!in_nullhost(inp->inp_laddr) &&
204 1.32 mycroft !in_hosteq(inp->inp_laddr, ip->ip_dst))
205 1.13 mycroft continue;
206 1.32 mycroft if (!in_nullhost(inp->inp_faddr) &&
207 1.32 mycroft !in_hosteq(inp->inp_faddr, ip->ip_src))
208 1.13 mycroft continue;
209 1.97 dyoung if (last == NULL)
210 1.97 dyoung ;
211 1.116 christos #if defined(IPSEC)
212 1.97 dyoung /* check AH/ESP integrity. */
213 1.124 christos else if (ipsec_used &&
214 1.124 christos ipsec4_in_reject_so(m, last->inp_socket)) {
215 1.106 thorpej IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
216 1.97 dyoung /* do not inject data to pcb */
217 1.97 dyoung }
218 1.97 dyoung #endif /*IPSEC*/
219 1.99 dyoung else if ((n = m_copypacket(m, M_DONTWAIT)) != NULL) {
220 1.100 dyoung rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts,
221 1.100 dyoung n);
222 1.97 dyoung opts = NULL;
223 1.13 mycroft }
224 1.36 thorpej last = inp;
225 1.13 mycroft }
226 1.116 christos #if defined(IPSEC)
227 1.55 itojun /* check AH/ESP integrity. */
228 1.124 christos if (ipsec_used && last != NULL
229 1.124 christos && ipsec4_in_reject_so(m, last->inp_socket)) {
230 1.55 itojun m_freem(m);
231 1.106 thorpej IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
232 1.105 thorpej IP_STATDEC(IP_STAT_DELIVERED);
233 1.55 itojun /* do not inject data to pcb */
234 1.55 itojun } else
235 1.55 itojun #endif /*IPSEC*/
236 1.100 dyoung if (last != NULL)
237 1.100 dyoung rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts, m);
238 1.100 dyoung else if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) {
239 1.105 thorpej uint64_t *ips;
240 1.105 thorpej
241 1.97 dyoung icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL,
242 1.97 dyoung 0, 0);
243 1.105 thorpej ips = IP_STAT_GETREF();
244 1.105 thorpej ips[IP_STAT_NOPROTO]++;
245 1.105 thorpej ips[IP_STAT_DELIVERED]--;
246 1.105 thorpej IP_STAT_PUTREF();
247 1.97 dyoung } else
248 1.97 dyoung m_freem(m);
249 1.43 itojun return;
250 1.60 itojun }
251 1.60 itojun
252 1.60 itojun int
253 1.83 perry rip_pcbnotify(struct inpcbtable *table,
254 1.83 perry struct in_addr faddr, struct in_addr laddr, int proto, int errno,
255 1.83 perry void (*notify)(struct inpcb *, int))
256 1.60 itojun {
257 1.117 christos struct inpcb_hdr *inph, *ninph;
258 1.60 itojun int nmatch;
259 1.60 itojun
260 1.60 itojun nmatch = 0;
261 1.117 christos TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) {
262 1.117 christos struct inpcb *inp = (struct inpcb *)inph;
263 1.75 itojun if (inp->inp_af != AF_INET)
264 1.75 itojun continue;
265 1.60 itojun if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
266 1.60 itojun continue;
267 1.60 itojun if (in_hosteq(inp->inp_faddr, faddr) &&
268 1.60 itojun in_hosteq(inp->inp_laddr, laddr)) {
269 1.60 itojun (*notify)(inp, errno);
270 1.60 itojun nmatch++;
271 1.60 itojun }
272 1.60 itojun }
273 1.60 itojun
274 1.60 itojun return nmatch;
275 1.60 itojun }
276 1.60 itojun
277 1.60 itojun void *
278 1.95 dyoung rip_ctlinput(int cmd, const struct sockaddr *sa, void *v)
279 1.60 itojun {
280 1.60 itojun struct ip *ip = v;
281 1.82 perry void (*notify)(struct inpcb *, int) = in_rtchange;
282 1.60 itojun int errno;
283 1.60 itojun
284 1.60 itojun if (sa->sa_family != AF_INET ||
285 1.60 itojun sa->sa_len != sizeof(struct sockaddr_in))
286 1.60 itojun return NULL;
287 1.60 itojun if ((unsigned)cmd >= PRC_NCMDS)
288 1.60 itojun return NULL;
289 1.60 itojun errno = inetctlerrmap[cmd];
290 1.60 itojun if (PRC_IS_REDIRECT(cmd))
291 1.60 itojun notify = in_rtchange, ip = 0;
292 1.60 itojun else if (cmd == PRC_HOSTDEAD)
293 1.60 itojun ip = 0;
294 1.60 itojun else if (errno == 0)
295 1.60 itojun return NULL;
296 1.60 itojun if (ip) {
297 1.95 dyoung rip_pcbnotify(&rawcbtable, satocsin(sa)->sin_addr,
298 1.60 itojun ip->ip_src, ip->ip_p, errno, notify);
299 1.60 itojun
300 1.60 itojun /* XXX mapped address case */
301 1.60 itojun } else
302 1.95 dyoung in_pcbnotifyall(&rawcbtable, satocsin(sa)->sin_addr, errno,
303 1.60 itojun notify);
304 1.60 itojun return NULL;
305 1.1 cgd }
306 1.1 cgd
307 1.1 cgd /*
308 1.1 cgd * Generate IP header and pass packet to ip_output.
309 1.1 cgd * Tack on options user may have setup with control call.
310 1.1 cgd */
311 1.9 mycroft int
312 1.24 christos rip_output(struct mbuf *m, ...)
313 1.24 christos {
314 1.53 augustss struct inpcb *inp;
315 1.53 augustss struct ip *ip;
316 1.10 mycroft struct mbuf *opts;
317 1.24 christos int flags;
318 1.24 christos va_list ap;
319 1.24 christos
320 1.24 christos va_start(ap, m);
321 1.27 mycroft inp = va_arg(ap, struct inpcb *);
322 1.24 christos va_end(ap);
323 1.24 christos
324 1.27 mycroft flags =
325 1.37 matt (inp->inp_socket->so_options & SO_DONTROUTE) | IP_ALLOWBROADCAST
326 1.37 matt | IP_RETURNMTU;
327 1.1 cgd
328 1.1 cgd /*
329 1.1 cgd * If the user handed us a complete IP packet, use it.
330 1.1 cgd * Otherwise, allocate an mbuf for a header and fill it in.
331 1.1 cgd */
332 1.13 mycroft if ((inp->inp_flags & INP_HDRINCL) == 0) {
333 1.35 thorpej if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
334 1.35 thorpej m_freem(m);
335 1.35 thorpej return (EMSGSIZE);
336 1.35 thorpej }
337 1.68 itojun M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
338 1.68 itojun if (!m)
339 1.68 itojun return (ENOBUFS);
340 1.1 cgd ip = mtod(m, struct ip *);
341 1.1 cgd ip->ip_tos = 0;
342 1.62 itojun ip->ip_off = htons(0);
343 1.13 mycroft ip->ip_p = inp->inp_ip.ip_p;
344 1.62 itojun ip->ip_len = htons(m->m_pkthdr.len);
345 1.13 mycroft ip->ip_src = inp->inp_laddr;
346 1.27 mycroft ip->ip_dst = inp->inp_faddr;
347 1.1 cgd ip->ip_ttl = MAXTTL;
348 1.13 mycroft opts = inp->inp_options;
349 1.13 mycroft } else {
350 1.35 thorpej if (m->m_pkthdr.len > IP_MAXPACKET) {
351 1.35 thorpej m_freem(m);
352 1.35 thorpej return (EMSGSIZE);
353 1.35 thorpej }
354 1.13 mycroft ip = mtod(m, struct ip *);
355 1.65 thorpej
356 1.65 thorpej /*
357 1.65 thorpej * If the mbuf is read-only, we need to allocate
358 1.65 thorpej * a new mbuf for the header, since we need to
359 1.65 thorpej * modify the header.
360 1.65 thorpej */
361 1.65 thorpej if (M_READONLY(m)) {
362 1.65 thorpej int hlen = ip->ip_hl << 2;
363 1.65 thorpej
364 1.65 thorpej m = m_copyup(m, hlen, (max_linkhdr + 3) & ~3);
365 1.65 thorpej if (m == NULL)
366 1.65 thorpej return (ENOMEM); /* XXX */
367 1.65 thorpej ip = mtod(m, struct ip *);
368 1.65 thorpej }
369 1.65 thorpej
370 1.62 itojun /* XXX userland passes ip_len and ip_off in host order */
371 1.38 mycroft if (m->m_pkthdr.len != ip->ip_len) {
372 1.38 mycroft m_freem(m);
373 1.38 mycroft return (EINVAL);
374 1.38 mycroft }
375 1.62 itojun HTONS(ip->ip_len);
376 1.62 itojun HTONS(ip->ip_off);
377 1.103 matt if (ip->ip_id != 0 || m->m_pkthdr.len < IP_MINFRAGSIZE)
378 1.103 matt flags |= IP_NOIPNEWID;
379 1.13 mycroft opts = NULL;
380 1.13 mycroft /* XXX prevent ip_output from overwriting header fields */
381 1.13 mycroft flags |= IP_RAWOUTPUT;
382 1.105 thorpej IP_STATINC(IP_STAT_RAWOUT);
383 1.1 cgd }
384 1.123 rmind
385 1.123 rmind /*
386 1.123 rmind * IP output. Note: if IP_RETURNMTU flag is set, the MTU size
387 1.123 rmind * will be stored in inp_errormtu.
388 1.123 rmind */
389 1.123 rmind return ip_output(m, opts, &inp->inp_route, flags, inp->inp_moptions,
390 1.123 rmind inp->inp_socket);
391 1.1 cgd }
392 1.1 cgd
393 1.1 cgd /*
394 1.1 cgd * Raw IP socket option processing.
395 1.1 cgd */
396 1.9 mycroft int
397 1.108 plunky rip_ctloutput(int op, struct socket *so, struct sockopt *sopt)
398 1.1 cgd {
399 1.53 augustss struct inpcb *inp = sotoinpcb(so);
400 1.31 mycroft int error = 0;
401 1.108 plunky int optval;
402 1.1 cgd
403 1.108 plunky if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) {
404 1.100 dyoung if (op == PRCO_GETOPT) {
405 1.108 plunky optval = (inp->inp_flags & INP_NOHEADER) ? 1 : 0;
406 1.108 plunky error = sockopt_set(sopt, &optval, sizeof(optval));
407 1.108 plunky } else if (op == PRCO_SETOPT) {
408 1.108 plunky error = sockopt_getint(sopt, &optval);
409 1.108 plunky if (error)
410 1.108 plunky goto out;
411 1.108 plunky if (optval) {
412 1.108 plunky inp->inp_flags &= ~INP_HDRINCL;
413 1.108 plunky inp->inp_flags |= INP_NOHEADER;
414 1.108 plunky } else
415 1.108 plunky inp->inp_flags &= ~INP_NOHEADER;
416 1.108 plunky }
417 1.108 plunky goto out;
418 1.108 plunky } else if (sopt->sopt_level != IPPROTO_IP)
419 1.108 plunky return ip_ctloutput(op, so, sopt);
420 1.100 dyoung
421 1.100 dyoung switch (op) {
422 1.31 mycroft
423 1.31 mycroft case PRCO_SETOPT:
424 1.108 plunky switch (sopt->sopt_name) {
425 1.31 mycroft case IP_HDRINCL:
426 1.108 plunky error = sockopt_getint(sopt, &optval);
427 1.108 plunky if (error)
428 1.108 plunky break;
429 1.108 plunky if (optval)
430 1.100 dyoung inp->inp_flags |= INP_HDRINCL;
431 1.100 dyoung else
432 1.100 dyoung inp->inp_flags &= ~INP_HDRINCL;
433 1.108 plunky break;
434 1.31 mycroft
435 1.31 mycroft #ifdef MROUTING
436 1.31 mycroft case MRT_INIT:
437 1.31 mycroft case MRT_DONE:
438 1.31 mycroft case MRT_ADD_VIF:
439 1.31 mycroft case MRT_DEL_VIF:
440 1.31 mycroft case MRT_ADD_MFC:
441 1.31 mycroft case MRT_DEL_MFC:
442 1.31 mycroft case MRT_ASSERT:
443 1.81 manu case MRT_API_CONFIG:
444 1.81 manu case MRT_ADD_BW_UPCALL:
445 1.81 manu case MRT_DEL_BW_UPCALL:
446 1.108 plunky error = ip_mrouter_set(so, sopt);
447 1.31 mycroft break;
448 1.31 mycroft #endif
449 1.31 mycroft
450 1.31 mycroft default:
451 1.108 plunky error = ip_ctloutput(op, so, sopt);
452 1.31 mycroft break;
453 1.13 mycroft }
454 1.13 mycroft break;
455 1.1 cgd
456 1.31 mycroft case PRCO_GETOPT:
457 1.108 plunky switch (sopt->sopt_name) {
458 1.31 mycroft case IP_HDRINCL:
459 1.108 plunky optval = inp->inp_flags & INP_HDRINCL;
460 1.108 plunky error = sockopt_set(sopt, &optval, sizeof(optval));
461 1.31 mycroft break;
462 1.31 mycroft
463 1.6 hpeyerl #ifdef MROUTING
464 1.31 mycroft case MRT_VERSION:
465 1.31 mycroft case MRT_ASSERT:
466 1.81 manu case MRT_API_SUPPORT:
467 1.81 manu case MRT_API_CONFIG:
468 1.108 plunky error = ip_mrouter_get(so, sopt);
469 1.18 mycroft break;
470 1.31 mycroft #endif
471 1.31 mycroft
472 1.18 mycroft default:
473 1.108 plunky error = ip_ctloutput(op, so, sopt);
474 1.18 mycroft break;
475 1.18 mycroft }
476 1.31 mycroft break;
477 1.1 cgd }
478 1.108 plunky out:
479 1.100 dyoung return error;
480 1.1 cgd }
481 1.1 cgd
482 1.27 mycroft int
483 1.137 rtr rip_connect_pcb(struct inpcb *inp, struct mbuf *nam)
484 1.27 mycroft {
485 1.27 mycroft struct sockaddr_in *addr = mtod(nam, struct sockaddr_in *);
486 1.27 mycroft
487 1.27 mycroft if (nam->m_len != sizeof(*addr))
488 1.27 mycroft return (EINVAL);
489 1.138 ozaki if (IFNET_EMPTY())
490 1.27 mycroft return (EADDRNOTAVAIL);
491 1.115 joerg if (addr->sin_family != AF_INET)
492 1.27 mycroft return (EAFNOSUPPORT);
493 1.27 mycroft inp->inp_faddr = addr->sin_addr;
494 1.27 mycroft return (0);
495 1.27 mycroft }
496 1.27 mycroft
497 1.139 rtr static void
498 1.139 rtr rip_disconnect1(struct inpcb *inp)
499 1.27 mycroft {
500 1.27 mycroft
501 1.32 mycroft inp->inp_faddr = zeroin_addr;
502 1.27 mycroft }
503 1.27 mycroft
504 1.121 rmind static int
505 1.121 rmind rip_attach(struct socket *so, int proto)
506 1.121 rmind {
507 1.121 rmind struct inpcb *inp;
508 1.121 rmind int error;
509 1.121 rmind
510 1.121 rmind KASSERT(sotoinpcb(so) == NULL);
511 1.121 rmind sosetlock(so);
512 1.121 rmind
513 1.121 rmind if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
514 1.121 rmind error = soreserve(so, rip_sendspace, rip_recvspace);
515 1.121 rmind if (error) {
516 1.121 rmind return error;
517 1.121 rmind }
518 1.121 rmind }
519 1.121 rmind
520 1.121 rmind error = in_pcballoc(so, &rawcbtable);
521 1.121 rmind if (error) {
522 1.121 rmind return error;
523 1.121 rmind }
524 1.121 rmind inp = sotoinpcb(so);
525 1.121 rmind inp->inp_ip.ip_p = proto;
526 1.121 rmind KASSERT(solocked(so));
527 1.121 rmind
528 1.121 rmind return 0;
529 1.121 rmind }
530 1.13 mycroft
531 1.121 rmind static void
532 1.121 rmind rip_detach(struct socket *so)
533 1.1 cgd {
534 1.53 augustss struct inpcb *inp;
535 1.121 rmind
536 1.121 rmind KASSERT(solocked(so));
537 1.121 rmind inp = sotoinpcb(so);
538 1.121 rmind KASSERT(inp != NULL);
539 1.121 rmind
540 1.13 mycroft #ifdef MROUTING
541 1.6 hpeyerl extern struct socket *ip_mrouter;
542 1.121 rmind if (so == ip_mrouter) {
543 1.121 rmind ip_mrouter_done();
544 1.121 rmind }
545 1.6 hpeyerl #endif
546 1.121 rmind in_pcbdetach(inp);
547 1.121 rmind }
548 1.121 rmind
549 1.125 rtr static int
550 1.133 rtr rip_accept(struct socket *so, struct mbuf *nam)
551 1.133 rtr {
552 1.133 rtr KASSERT(solocked(so));
553 1.133 rtr
554 1.133 rtr panic("rip_accept");
555 1.136 rtr
556 1.136 rtr return EOPNOTSUPP;
557 1.136 rtr }
558 1.136 rtr
559 1.136 rtr static int
560 1.142 rtr rip_bind(struct socket *so, struct mbuf *nam, struct lwp *l)
561 1.136 rtr {
562 1.136 rtr struct inpcb *inp = sotoinpcb(so);
563 1.136 rtr struct sockaddr_in *addr;
564 1.136 rtr int error = 0;
565 1.136 rtr int s;
566 1.136 rtr
567 1.136 rtr KASSERT(solocked(so));
568 1.136 rtr KASSERT(inp != NULL);
569 1.136 rtr KASSERT(nam != NULL);
570 1.136 rtr
571 1.136 rtr s = splsoftnet();
572 1.136 rtr addr = mtod(nam, struct sockaddr_in *);
573 1.136 rtr if (nam->m_len != sizeof(*addr)) {
574 1.136 rtr error = EINVAL;
575 1.136 rtr goto release;
576 1.136 rtr }
577 1.138 ozaki if (IFNET_EMPTY()) {
578 1.136 rtr error = EADDRNOTAVAIL;
579 1.136 rtr goto release;
580 1.136 rtr }
581 1.136 rtr if (addr->sin_family != AF_INET) {
582 1.136 rtr error = EAFNOSUPPORT;
583 1.136 rtr goto release;
584 1.136 rtr }
585 1.136 rtr if (!in_nullhost(addr->sin_addr) &&
586 1.136 rtr ifa_ifwithaddr(sintosa(addr)) == 0) {
587 1.136 rtr error = EADDRNOTAVAIL;
588 1.136 rtr goto release;
589 1.136 rtr }
590 1.136 rtr inp->inp_laddr = addr->sin_addr;
591 1.136 rtr
592 1.136 rtr release:
593 1.136 rtr splx(s);
594 1.136 rtr return error;
595 1.136 rtr }
596 1.136 rtr
597 1.136 rtr static int
598 1.142 rtr rip_listen(struct socket *so, struct lwp *l)
599 1.136 rtr {
600 1.136 rtr KASSERT(solocked(so));
601 1.136 rtr
602 1.133 rtr return EOPNOTSUPP;
603 1.133 rtr }
604 1.133 rtr
605 1.133 rtr static int
606 1.142 rtr rip_connect(struct socket *so, struct mbuf *nam, struct lwp *l)
607 1.137 rtr {
608 1.137 rtr struct inpcb *inp = sotoinpcb(so);
609 1.137 rtr int error = 0;
610 1.137 rtr int s;
611 1.137 rtr
612 1.137 rtr KASSERT(solocked(so));
613 1.137 rtr KASSERT(inp != NULL);
614 1.137 rtr KASSERT(nam != NULL);
615 1.137 rtr
616 1.137 rtr s = splsoftnet();
617 1.137 rtr error = rip_connect_pcb(inp, nam);
618 1.137 rtr if (! error)
619 1.137 rtr soisconnected(so);
620 1.140 rtr splx(s);
621 1.137 rtr
622 1.137 rtr return error;
623 1.137 rtr }
624 1.137 rtr
625 1.139 rtr static int
626 1.145 rtr rip_connect2(struct socket *so, struct socket *so2)
627 1.145 rtr {
628 1.145 rtr KASSERT(solocked(so));
629 1.145 rtr
630 1.145 rtr return EOPNOTSUPP;
631 1.145 rtr }
632 1.145 rtr
633 1.145 rtr static int
634 1.139 rtr rip_disconnect(struct socket *so)
635 1.139 rtr {
636 1.139 rtr struct inpcb *inp = sotoinpcb(so);
637 1.140 rtr int s;
638 1.139 rtr
639 1.139 rtr KASSERT(solocked(so));
640 1.139 rtr KASSERT(inp != NULL);
641 1.139 rtr
642 1.140 rtr s = splsoftnet();
643 1.139 rtr soisdisconnected(so);
644 1.139 rtr rip_disconnect1(inp);
645 1.140 rtr splx(s);
646 1.140 rtr
647 1.139 rtr return 0;
648 1.139 rtr }
649 1.139 rtr
650 1.139 rtr static int
651 1.139 rtr rip_shutdown(struct socket *so)
652 1.139 rtr {
653 1.140 rtr int s;
654 1.140 rtr
655 1.139 rtr KASSERT(solocked(so));
656 1.139 rtr
657 1.139 rtr /*
658 1.139 rtr * Mark the connection as being incapable of further input.
659 1.139 rtr */
660 1.140 rtr s = splsoftnet();
661 1.139 rtr socantsendmore(so);
662 1.140 rtr splx(s);
663 1.140 rtr
664 1.139 rtr return 0;
665 1.139 rtr }
666 1.139 rtr
667 1.139 rtr static int
668 1.139 rtr rip_abort(struct socket *so)
669 1.139 rtr {
670 1.139 rtr KASSERT(solocked(so));
671 1.139 rtr
672 1.139 rtr panic("rip_abort");
673 1.139 rtr
674 1.139 rtr return EOPNOTSUPP;
675 1.139 rtr }
676 1.137 rtr
677 1.137 rtr static int
678 1.127 rtr rip_ioctl(struct socket *so, u_long cmd, void *nam, struct ifnet *ifp)
679 1.125 rtr {
680 1.127 rtr return in_control(so, cmd, nam, ifp);
681 1.125 rtr }
682 1.125 rtr
683 1.128 rtr static int
684 1.128 rtr rip_stat(struct socket *so, struct stat *ub)
685 1.128 rtr {
686 1.131 rtr KASSERT(solocked(so));
687 1.131 rtr
688 1.130 rtr /* stat: don't bother with a blocksize. */
689 1.130 rtr return 0;
690 1.128 rtr }
691 1.128 rtr
692 1.132 rtr static int
693 1.132 rtr rip_peeraddr(struct socket *so, struct mbuf *nam)
694 1.132 rtr {
695 1.140 rtr int s;
696 1.140 rtr
697 1.134 rtr KASSERT(solocked(so));
698 1.132 rtr KASSERT(sotoinpcb(so) != NULL);
699 1.132 rtr KASSERT(nam != NULL);
700 1.132 rtr
701 1.140 rtr s = splsoftnet();
702 1.132 rtr in_setpeeraddr(sotoinpcb(so), nam);
703 1.140 rtr splx(s);
704 1.140 rtr
705 1.132 rtr return 0;
706 1.132 rtr }
707 1.132 rtr
708 1.132 rtr static int
709 1.132 rtr rip_sockaddr(struct socket *so, struct mbuf *nam)
710 1.132 rtr {
711 1.140 rtr int s;
712 1.140 rtr
713 1.134 rtr KASSERT(solocked(so));
714 1.132 rtr KASSERT(sotoinpcb(so) != NULL);
715 1.132 rtr KASSERT(nam != NULL);
716 1.132 rtr
717 1.140 rtr s = splsoftnet();
718 1.132 rtr in_setsockaddr(sotoinpcb(so), nam);
719 1.140 rtr splx(s);
720 1.140 rtr
721 1.132 rtr return 0;
722 1.132 rtr }
723 1.132 rtr
724 1.135 rtr static int
725 1.144 rtr rip_rcvd(struct socket *so, int flags, struct lwp *l)
726 1.144 rtr {
727 1.144 rtr KASSERT(solocked(so));
728 1.144 rtr
729 1.144 rtr return EOPNOTSUPP;
730 1.144 rtr }
731 1.144 rtr
732 1.144 rtr static int
733 1.135 rtr rip_recvoob(struct socket *so, struct mbuf *m, int flags)
734 1.135 rtr {
735 1.135 rtr KASSERT(solocked(so));
736 1.135 rtr
737 1.135 rtr return EOPNOTSUPP;
738 1.135 rtr }
739 1.135 rtr
740 1.135 rtr static int
741 1.143 rtr rip_send(struct socket *so, struct mbuf *m, struct mbuf *nam,
742 1.143 rtr struct mbuf *control, struct lwp *l)
743 1.143 rtr {
744 1.143 rtr struct inpcb *inp = sotoinpcb(so);
745 1.143 rtr int error = 0;
746 1.143 rtr int s;
747 1.143 rtr
748 1.143 rtr KASSERT(solocked(so));
749 1.143 rtr KASSERT(inp != NULL);
750 1.143 rtr KASSERT(m != NULL);
751 1.143 rtr
752 1.143 rtr /*
753 1.143 rtr * Ship a packet out. The appropriate raw output
754 1.143 rtr * routine handles any massaging necessary.
755 1.143 rtr */
756 1.143 rtr if (control && control->m_len) {
757 1.143 rtr m_freem(control);
758 1.143 rtr m_freem(m);
759 1.143 rtr return EINVAL;
760 1.143 rtr }
761 1.143 rtr
762 1.143 rtr s = splsoftnet();
763 1.143 rtr if (nam) {
764 1.143 rtr if ((so->so_state & SS_ISCONNECTED) != 0) {
765 1.143 rtr error = EISCONN;
766 1.143 rtr goto die;
767 1.143 rtr }
768 1.143 rtr error = rip_connect_pcb(inp, nam);
769 1.143 rtr if (error) {
770 1.143 rtr die:
771 1.143 rtr m_freem(m);
772 1.143 rtr splx(s);
773 1.143 rtr return error;
774 1.143 rtr }
775 1.143 rtr } else {
776 1.143 rtr if ((so->so_state & SS_ISCONNECTED) == 0) {
777 1.143 rtr error = ENOTCONN;
778 1.143 rtr goto die;
779 1.143 rtr }
780 1.143 rtr }
781 1.143 rtr error = rip_output(m, inp);
782 1.143 rtr if (nam)
783 1.143 rtr rip_disconnect1(inp);
784 1.143 rtr
785 1.143 rtr splx(s);
786 1.143 rtr return error;
787 1.143 rtr }
788 1.143 rtr
789 1.143 rtr static int
790 1.135 rtr rip_sendoob(struct socket *so, struct mbuf *m, struct mbuf *control)
791 1.135 rtr {
792 1.135 rtr KASSERT(solocked(so));
793 1.135 rtr
794 1.135 rtr m_freem(m);
795 1.135 rtr m_freem(control);
796 1.135 rtr
797 1.135 rtr return EOPNOTSUPP;
798 1.135 rtr }
799 1.135 rtr
800 1.145 rtr static int
801 1.145 rtr rip_purgeif(struct socket *so, struct ifnet *ifp)
802 1.145 rtr {
803 1.145 rtr int s;
804 1.145 rtr
805 1.145 rtr s = splsoftnet();
806 1.145 rtr mutex_enter(softnet_lock);
807 1.145 rtr in_pcbpurgeif0(&rawcbtable, ifp);
808 1.145 rtr in_purgeif(ifp);
809 1.145 rtr in_pcbpurgeif(&rawcbtable, ifp);
810 1.145 rtr mutex_exit(softnet_lock);
811 1.145 rtr splx(s);
812 1.145 rtr
813 1.145 rtr return 0;
814 1.145 rtr }
815 1.145 rtr
816 1.121 rmind int
817 1.121 rmind rip_usrreq(struct socket *so, int req, struct mbuf *m, struct mbuf *nam,
818 1.121 rmind struct mbuf *control, struct lwp *l)
819 1.121 rmind {
820 1.121 rmind KASSERT(req != PRU_ATTACH);
821 1.121 rmind KASSERT(req != PRU_DETACH);
822 1.133 rtr KASSERT(req != PRU_ACCEPT);
823 1.136 rtr KASSERT(req != PRU_BIND);
824 1.136 rtr KASSERT(req != PRU_LISTEN);
825 1.137 rtr KASSERT(req != PRU_CONNECT);
826 1.145 rtr KASSERT(req != PRU_CONNECT2);
827 1.139 rtr KASSERT(req != PRU_DISCONNECT);
828 1.139 rtr KASSERT(req != PRU_SHUTDOWN);
829 1.139 rtr KASSERT(req != PRU_ABORT);
830 1.125 rtr KASSERT(req != PRU_CONTROL);
831 1.128 rtr KASSERT(req != PRU_SENSE);
832 1.132 rtr KASSERT(req != PRU_PEERADDR);
833 1.132 rtr KASSERT(req != PRU_SOCKADDR);
834 1.144 rtr KASSERT(req != PRU_RCVD);
835 1.135 rtr KASSERT(req != PRU_RCVOOB);
836 1.143 rtr KASSERT(req != PRU_SEND);
837 1.135 rtr KASSERT(req != PRU_SENDOOB);
838 1.145 rtr KASSERT(req != PRU_PURGEIF);
839 1.22 pk
840 1.121 rmind KASSERT(solocked(so));
841 1.121 rmind
842 1.145 rtr if (sotoinpcb(so) == NULL)
843 1.121 rmind return EINVAL;
844 1.22 pk
845 1.145 rtr panic("rip_usrreq");
846 1.1 cgd
847 1.145 rtr return 0;
848 1.1 cgd }
849 1.84 atatat
850 1.122 rmind PR_WRAP_USRREQS(rip)
851 1.122 rmind #define rip_attach rip_attach_wrapper
852 1.122 rmind #define rip_detach rip_detach_wrapper
853 1.133 rtr #define rip_accept rip_accept_wrapper
854 1.136 rtr #define rip_bind rip_bind_wrapper
855 1.136 rtr #define rip_listen rip_listen_wrapper
856 1.137 rtr #define rip_connect rip_connect_wrapper
857 1.145 rtr #define rip_connect2 rip_connect2_wrapper
858 1.139 rtr #define rip_disconnect rip_disconnect_wrapper
859 1.139 rtr #define rip_shutdown rip_shutdown_wrapper
860 1.139 rtr #define rip_abort rip_abort_wrapper
861 1.125 rtr #define rip_ioctl rip_ioctl_wrapper
862 1.128 rtr #define rip_stat rip_stat_wrapper
863 1.132 rtr #define rip_peeraddr rip_peeraddr_wrapper
864 1.132 rtr #define rip_sockaddr rip_sockaddr_wrapper
865 1.144 rtr #define rip_rcvd rip_rcvd_wrapper
866 1.135 rtr #define rip_recvoob rip_recvoob_wrapper
867 1.143 rtr #define rip_send rip_send_wrapper
868 1.135 rtr #define rip_sendoob rip_sendoob_wrapper
869 1.145 rtr #define rip_purgeif rip_purgeif_wrapper
870 1.120 rmind #define rip_usrreq rip_usrreq_wrapper
871 1.120 rmind
872 1.120 rmind const struct pr_usrreqs rip_usrreqs = {
873 1.121 rmind .pr_attach = rip_attach,
874 1.121 rmind .pr_detach = rip_detach,
875 1.133 rtr .pr_accept = rip_accept,
876 1.136 rtr .pr_bind = rip_bind,
877 1.136 rtr .pr_listen = rip_listen,
878 1.137 rtr .pr_connect = rip_connect,
879 1.145 rtr .pr_connect2 = rip_connect2,
880 1.139 rtr .pr_disconnect = rip_disconnect,
881 1.139 rtr .pr_shutdown = rip_shutdown,
882 1.139 rtr .pr_abort = rip_abort,
883 1.125 rtr .pr_ioctl = rip_ioctl,
884 1.128 rtr .pr_stat = rip_stat,
885 1.132 rtr .pr_peeraddr = rip_peeraddr,
886 1.132 rtr .pr_sockaddr = rip_sockaddr,
887 1.144 rtr .pr_rcvd = rip_rcvd,
888 1.135 rtr .pr_recvoob = rip_recvoob,
889 1.143 rtr .pr_send = rip_send,
890 1.135 rtr .pr_sendoob = rip_sendoob,
891 1.145 rtr .pr_purgeif = rip_purgeif,
892 1.120 rmind .pr_generic = rip_usrreq,
893 1.120 rmind };
894 1.120 rmind
895 1.110 pooka static void
896 1.110 pooka sysctl_net_inet_raw_setup(struct sysctllog **clog)
897 1.84 atatat {
898 1.84 atatat
899 1.84 atatat sysctl_createv(clog, 0, NULL, NULL,
900 1.84 atatat CTLFLAG_PERMANENT,
901 1.84 atatat CTLTYPE_NODE, "inet", NULL,
902 1.84 atatat NULL, 0, NULL, 0,
903 1.84 atatat CTL_NET, PF_INET, CTL_EOL);
904 1.84 atatat sysctl_createv(clog, 0, NULL, NULL,
905 1.84 atatat CTLFLAG_PERMANENT,
906 1.84 atatat CTLTYPE_NODE, "raw",
907 1.84 atatat SYSCTL_DESCR("Raw IPv4 settings"),
908 1.84 atatat NULL, 0, NULL, 0,
909 1.84 atatat CTL_NET, PF_INET, IPPROTO_RAW, CTL_EOL);
910 1.84 atatat
911 1.84 atatat sysctl_createv(clog, 0, NULL, NULL,
912 1.84 atatat CTLFLAG_PERMANENT,
913 1.86 atatat CTLTYPE_STRUCT, "pcblist",
914 1.84 atatat SYSCTL_DESCR("Raw IPv4 control block list"),
915 1.84 atatat sysctl_inpcblist, 0, &rawcbtable, 0,
916 1.84 atatat CTL_NET, PF_INET, IPPROTO_RAW,
917 1.84 atatat CTL_CREATE, CTL_EOL);
918 1.84 atatat }
919