raw_ip.c revision 1.169 1 1.169 maxv /* $NetBSD: raw_ip.c,v 1.169 2018/02/26 09:04:29 maxv 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 1.169 maxv __KERNEL_RCSID(0, "$NetBSD: raw_ip.c,v 1.169 2018/02/26 09:04:29 maxv Exp $");
69 1.40 scottr
70 1.153 pooka #ifdef _KERNEL_OPT
71 1.78 jonathan #include "opt_inet.h"
72 1.45 thorpej #include "opt_ipsec.h"
73 1.40 scottr #include "opt_mrouting.h"
74 1.162 ozaki #include "opt_net_mpsafe.h"
75 1.153 pooka #endif
76 1.1 cgd
77 1.7 mycroft #include <sys/param.h>
78 1.84 atatat #include <sys/sysctl.h>
79 1.7 mycroft #include <sys/mbuf.h>
80 1.7 mycroft #include <sys/socket.h>
81 1.7 mycroft #include <sys/protosw.h>
82 1.7 mycroft #include <sys/socketvar.h>
83 1.7 mycroft #include <sys/errno.h>
84 1.13 mycroft #include <sys/systm.h>
85 1.26 mycroft #include <sys/proc.h>
86 1.89 elad #include <sys/kauth.h>
87 1.1 cgd
88 1.7 mycroft #include <net/if.h>
89 1.1 cgd
90 1.7 mycroft #include <netinet/in.h>
91 1.7 mycroft #include <netinet/in_systm.h>
92 1.7 mycroft #include <netinet/ip.h>
93 1.7 mycroft #include <netinet/ip_var.h>
94 1.105 thorpej #include <netinet/ip_private.h>
95 1.13 mycroft #include <netinet/ip_mroute.h>
96 1.44 darrenr #include <netinet/ip_icmp.h>
97 1.7 mycroft #include <netinet/in_pcb.h>
98 1.87 yamt #include <netinet/in_proto.h>
99 1.24 christos #include <netinet/in_var.h>
100 1.24 christos
101 1.116 christos #ifdef IPSEC
102 1.72 jonathan #include <netipsec/ipsec.h>
103 1.106 thorpej #include <netipsec/ipsec_var.h>
104 1.106 thorpej #include <netipsec/ipsec_private.h>
105 1.116 christos #endif /* IPSEC */
106 1.72 jonathan
107 1.20 mycroft struct inpcbtable rawcbtable;
108 1.30 pk
109 1.82 perry int rip_pcbnotify(struct inpcbtable *, struct in_addr,
110 1.82 perry struct in_addr, int, int, void (*)(struct inpcb *, int));
111 1.149 rtr static int rip_connect_pcb(struct inpcb *, struct sockaddr_in *);
112 1.139 rtr static void rip_disconnect1(struct inpcb *);
113 1.13 mycroft
114 1.110 pooka static void sysctl_net_inet_raw_setup(struct sysctllog **);
115 1.110 pooka
116 1.13 mycroft /*
117 1.13 mycroft * Nominal space allocated to a raw ip socket.
118 1.13 mycroft */
119 1.13 mycroft #define RIPSNDQ 8192
120 1.13 mycroft #define RIPRCVQ 8192
121 1.1 cgd
122 1.121 rmind static u_long rip_sendspace = RIPSNDQ;
123 1.121 rmind static u_long rip_recvspace = RIPRCVQ;
124 1.121 rmind
125 1.1 cgd /*
126 1.1 cgd * Raw interface to IP protocol.
127 1.1 cgd */
128 1.13 mycroft
129 1.13 mycroft /*
130 1.13 mycroft * Initialize raw connection block q.
131 1.13 mycroft */
132 1.13 mycroft void
133 1.83 perry rip_init(void)
134 1.13 mycroft {
135 1.13 mycroft
136 1.110 pooka sysctl_net_inet_raw_setup(NULL);
137 1.33 mycroft in_pcbinit(&rawcbtable, 1, 1);
138 1.13 mycroft }
139 1.13 mycroft
140 1.100 dyoung static void
141 1.100 dyoung rip_sbappendaddr(struct inpcb *last, struct ip *ip, const struct sockaddr *sa,
142 1.100 dyoung int hlen, struct mbuf *opts, struct mbuf *n)
143 1.100 dyoung {
144 1.100 dyoung if (last->inp_flags & INP_NOHEADER)
145 1.100 dyoung m_adj(n, hlen);
146 1.109 christos if (last->inp_flags & INP_CONTROLOPTS
147 1.165 christos || SOOPT_TIMESTAMP(last->inp_socket->so_options))
148 1.100 dyoung ip_savecontrol(last, &opts, ip, n);
149 1.100 dyoung if (sbappendaddr(&last->inp_socket->so_rcv, sa, n, opts) == 0) {
150 1.100 dyoung /* should notify about lost packet */
151 1.100 dyoung m_freem(n);
152 1.100 dyoung if (opts)
153 1.100 dyoung m_freem(opts);
154 1.100 dyoung } else
155 1.100 dyoung sorwakeup(last->inp_socket);
156 1.100 dyoung }
157 1.100 dyoung
158 1.1 cgd /*
159 1.1 cgd * Setup generic address and protocol structures
160 1.1 cgd * for raw_input routine, then pass them along with
161 1.1 cgd * mbuf chain.
162 1.1 cgd */
163 1.9 mycroft void
164 1.24 christos rip_input(struct mbuf *m, ...)
165 1.1 cgd {
166 1.100 dyoung int hlen, proto;
167 1.53 augustss struct ip *ip = mtod(m, struct ip *);
168 1.75 itojun struct inpcb_hdr *inph;
169 1.53 augustss struct inpcb *inp;
170 1.97 dyoung struct inpcb *last = NULL;
171 1.97 dyoung struct mbuf *n, *opts = NULL;
172 1.32 mycroft struct sockaddr_in ripsrc;
173 1.43 itojun va_list ap;
174 1.43 itojun
175 1.43 itojun va_start(ap, m);
176 1.64 simonb (void)va_arg(ap, int); /* ignore value, advance ap */
177 1.43 itojun proto = va_arg(ap, int);
178 1.43 itojun va_end(ap);
179 1.1 cgd
180 1.97 dyoung sockaddr_in_init(&ripsrc, &ip->ip_src, 0);
181 1.42 thorpej
182 1.42 thorpej /*
183 1.42 thorpej * XXX Compatibility: programs using raw IP expect ip_len
184 1.62 itojun * XXX to have the header length subtracted, and in host order.
185 1.62 itojun * XXX ip_off is also expected to be host order.
186 1.42 thorpej */
187 1.100 dyoung hlen = ip->ip_hl << 2;
188 1.100 dyoung ip->ip_len = ntohs(ip->ip_len) - hlen;
189 1.62 itojun NTOHS(ip->ip_off);
190 1.32 mycroft
191 1.117 christos TAILQ_FOREACH(inph, &rawcbtable.inpt_queue, inph_queue) {
192 1.75 itojun inp = (struct inpcb *)inph;
193 1.75 itojun if (inp->inp_af != AF_INET)
194 1.75 itojun continue;
195 1.43 itojun if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
196 1.13 mycroft continue;
197 1.32 mycroft if (!in_nullhost(inp->inp_laddr) &&
198 1.32 mycroft !in_hosteq(inp->inp_laddr, ip->ip_dst))
199 1.13 mycroft continue;
200 1.32 mycroft if (!in_nullhost(inp->inp_faddr) &&
201 1.32 mycroft !in_hosteq(inp->inp_faddr, ip->ip_src))
202 1.13 mycroft continue;
203 1.97 dyoung if (last == NULL)
204 1.97 dyoung ;
205 1.116 christos #if defined(IPSEC)
206 1.97 dyoung /* check AH/ESP integrity. */
207 1.169 maxv else if (ipsec_used && ipsec_in_reject(m, last)) {
208 1.106 thorpej IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
209 1.97 dyoung /* do not inject data to pcb */
210 1.97 dyoung }
211 1.97 dyoung #endif /*IPSEC*/
212 1.99 dyoung else if ((n = m_copypacket(m, M_DONTWAIT)) != NULL) {
213 1.100 dyoung rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts,
214 1.100 dyoung n);
215 1.97 dyoung opts = NULL;
216 1.13 mycroft }
217 1.36 thorpej last = inp;
218 1.13 mycroft }
219 1.116 christos #if defined(IPSEC)
220 1.55 itojun /* check AH/ESP integrity. */
221 1.169 maxv if (ipsec_used && last != NULL && ipsec_in_reject(m, last)) {
222 1.55 itojun m_freem(m);
223 1.106 thorpej IPSEC_STATINC(IPSEC_STAT_IN_POLVIO);
224 1.105 thorpej IP_STATDEC(IP_STAT_DELIVERED);
225 1.55 itojun /* do not inject data to pcb */
226 1.55 itojun } else
227 1.55 itojun #endif /*IPSEC*/
228 1.100 dyoung if (last != NULL)
229 1.100 dyoung rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts, m);
230 1.156 riastrad else if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) {
231 1.105 thorpej uint64_t *ips;
232 1.105 thorpej
233 1.97 dyoung icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL,
234 1.97 dyoung 0, 0);
235 1.105 thorpej ips = IP_STAT_GETREF();
236 1.105 thorpej ips[IP_STAT_NOPROTO]++;
237 1.105 thorpej ips[IP_STAT_DELIVERED]--;
238 1.105 thorpej IP_STAT_PUTREF();
239 1.97 dyoung } else
240 1.97 dyoung m_freem(m);
241 1.43 itojun return;
242 1.60 itojun }
243 1.60 itojun
244 1.60 itojun int
245 1.83 perry rip_pcbnotify(struct inpcbtable *table,
246 1.83 perry struct in_addr faddr, struct in_addr laddr, int proto, int errno,
247 1.83 perry void (*notify)(struct inpcb *, int))
248 1.60 itojun {
249 1.117 christos struct inpcb_hdr *inph, *ninph;
250 1.60 itojun int nmatch;
251 1.60 itojun
252 1.60 itojun nmatch = 0;
253 1.117 christos TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) {
254 1.117 christos struct inpcb *inp = (struct inpcb *)inph;
255 1.75 itojun if (inp->inp_af != AF_INET)
256 1.75 itojun continue;
257 1.60 itojun if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
258 1.60 itojun continue;
259 1.60 itojun if (in_hosteq(inp->inp_faddr, faddr) &&
260 1.60 itojun in_hosteq(inp->inp_laddr, laddr)) {
261 1.60 itojun (*notify)(inp, errno);
262 1.60 itojun nmatch++;
263 1.60 itojun }
264 1.60 itojun }
265 1.60 itojun
266 1.60 itojun return nmatch;
267 1.60 itojun }
268 1.60 itojun
269 1.60 itojun void *
270 1.95 dyoung rip_ctlinput(int cmd, const struct sockaddr *sa, void *v)
271 1.60 itojun {
272 1.60 itojun struct ip *ip = v;
273 1.82 perry void (*notify)(struct inpcb *, int) = in_rtchange;
274 1.60 itojun int errno;
275 1.60 itojun
276 1.60 itojun if (sa->sa_family != AF_INET ||
277 1.60 itojun sa->sa_len != sizeof(struct sockaddr_in))
278 1.60 itojun return NULL;
279 1.60 itojun if ((unsigned)cmd >= PRC_NCMDS)
280 1.60 itojun return NULL;
281 1.60 itojun errno = inetctlerrmap[cmd];
282 1.60 itojun if (PRC_IS_REDIRECT(cmd))
283 1.60 itojun notify = in_rtchange, ip = 0;
284 1.60 itojun else if (cmd == PRC_HOSTDEAD)
285 1.60 itojun ip = 0;
286 1.60 itojun else if (errno == 0)
287 1.60 itojun return NULL;
288 1.60 itojun if (ip) {
289 1.95 dyoung rip_pcbnotify(&rawcbtable, satocsin(sa)->sin_addr,
290 1.60 itojun ip->ip_src, ip->ip_p, errno, notify);
291 1.60 itojun
292 1.60 itojun /* XXX mapped address case */
293 1.60 itojun } else
294 1.95 dyoung in_pcbnotifyall(&rawcbtable, satocsin(sa)->sin_addr, errno,
295 1.60 itojun notify);
296 1.60 itojun return NULL;
297 1.1 cgd }
298 1.1 cgd
299 1.1 cgd /*
300 1.1 cgd * Generate IP header and pass packet to ip_output.
301 1.1 cgd * Tack on options user may have setup with control call.
302 1.1 cgd */
303 1.9 mycroft int
304 1.166 ryo rip_output(struct mbuf *m, struct inpcb *inp, struct mbuf *control,
305 1.166 ryo struct lwp *l)
306 1.24 christos {
307 1.53 augustss struct ip *ip;
308 1.10 mycroft struct mbuf *opts;
309 1.166 ryo struct ip_pktopts pktopts;
310 1.166 ryo kauth_cred_t cred;
311 1.166 ryo int error, flags;
312 1.166 ryo
313 1.166 ryo flags = (inp->inp_socket->so_options & SO_DONTROUTE) |
314 1.166 ryo IP_ALLOWBROADCAST | IP_RETURNMTU;
315 1.166 ryo
316 1.166 ryo if (l == NULL)
317 1.166 ryo cred = NULL;
318 1.166 ryo else
319 1.166 ryo cred = l->l_cred;
320 1.166 ryo
321 1.166 ryo /* Setup IP outgoing packet options */
322 1.166 ryo memset(&pktopts, 0, sizeof(pktopts));
323 1.167 ryo error = ip_setpktopts(control, &pktopts, &flags, inp, cred);
324 1.166 ryo if (control != NULL)
325 1.166 ryo m_freem(control);
326 1.166 ryo if (error != 0)
327 1.166 ryo goto release;
328 1.1 cgd
329 1.1 cgd /*
330 1.1 cgd * If the user handed us a complete IP packet, use it.
331 1.1 cgd * Otherwise, allocate an mbuf for a header and fill it in.
332 1.1 cgd */
333 1.13 mycroft if ((inp->inp_flags & INP_HDRINCL) == 0) {
334 1.35 thorpej if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
335 1.166 ryo error = EMSGSIZE;
336 1.166 ryo goto release;
337 1.35 thorpej }
338 1.68 itojun M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
339 1.166 ryo if (!m) {
340 1.166 ryo error = ENOBUFS;
341 1.166 ryo goto release;
342 1.166 ryo }
343 1.1 cgd ip = mtod(m, struct ip *);
344 1.1 cgd ip->ip_tos = 0;
345 1.62 itojun ip->ip_off = htons(0);
346 1.13 mycroft ip->ip_p = inp->inp_ip.ip_p;
347 1.62 itojun ip->ip_len = htons(m->m_pkthdr.len);
348 1.166 ryo ip->ip_src = pktopts.ippo_laddr.sin_addr;
349 1.27 mycroft ip->ip_dst = inp->inp_faddr;
350 1.1 cgd ip->ip_ttl = MAXTTL;
351 1.13 mycroft opts = inp->inp_options;
352 1.13 mycroft } else {
353 1.35 thorpej if (m->m_pkthdr.len > IP_MAXPACKET) {
354 1.166 ryo error = EMSGSIZE;
355 1.166 ryo goto release;
356 1.35 thorpej }
357 1.13 mycroft ip = mtod(m, struct ip *);
358 1.65 thorpej
359 1.65 thorpej /*
360 1.65 thorpej * If the mbuf is read-only, we need to allocate
361 1.65 thorpej * a new mbuf for the header, since we need to
362 1.65 thorpej * modify the header.
363 1.65 thorpej */
364 1.65 thorpej if (M_READONLY(m)) {
365 1.65 thorpej int hlen = ip->ip_hl << 2;
366 1.65 thorpej
367 1.65 thorpej m = m_copyup(m, hlen, (max_linkhdr + 3) & ~3);
368 1.166 ryo if (m == NULL) {
369 1.166 ryo error = ENOMEM; /* XXX */
370 1.166 ryo goto release;
371 1.166 ryo }
372 1.65 thorpej ip = mtod(m, struct ip *);
373 1.65 thorpej }
374 1.65 thorpej
375 1.62 itojun /* XXX userland passes ip_len and ip_off in host order */
376 1.38 mycroft if (m->m_pkthdr.len != ip->ip_len) {
377 1.166 ryo error = EINVAL;
378 1.166 ryo goto release;
379 1.38 mycroft }
380 1.62 itojun HTONS(ip->ip_len);
381 1.62 itojun HTONS(ip->ip_off);
382 1.103 matt if (ip->ip_id != 0 || m->m_pkthdr.len < IP_MINFRAGSIZE)
383 1.103 matt flags |= IP_NOIPNEWID;
384 1.13 mycroft opts = NULL;
385 1.13 mycroft /* XXX prevent ip_output from overwriting header fields */
386 1.13 mycroft flags |= IP_RAWOUTPUT;
387 1.105 thorpej IP_STATINC(IP_STAT_RAWOUT);
388 1.1 cgd }
389 1.123 rmind
390 1.123 rmind /*
391 1.123 rmind * IP output. Note: if IP_RETURNMTU flag is set, the MTU size
392 1.123 rmind * will be stored in inp_errormtu.
393 1.123 rmind */
394 1.166 ryo return ip_output(m, opts, &inp->inp_route, flags, pktopts.ippo_imo,
395 1.166 ryo inp);
396 1.166 ryo
397 1.166 ryo release:
398 1.166 ryo if (m != NULL)
399 1.166 ryo m_freem(m);
400 1.166 ryo return error;
401 1.1 cgd }
402 1.1 cgd
403 1.1 cgd /*
404 1.1 cgd * Raw IP socket option processing.
405 1.1 cgd */
406 1.9 mycroft int
407 1.108 plunky rip_ctloutput(int op, struct socket *so, struct sockopt *sopt)
408 1.1 cgd {
409 1.53 augustss struct inpcb *inp = sotoinpcb(so);
410 1.31 mycroft int error = 0;
411 1.108 plunky int optval;
412 1.1 cgd
413 1.108 plunky if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) {
414 1.100 dyoung if (op == PRCO_GETOPT) {
415 1.108 plunky optval = (inp->inp_flags & INP_NOHEADER) ? 1 : 0;
416 1.108 plunky error = sockopt_set(sopt, &optval, sizeof(optval));
417 1.108 plunky } else if (op == PRCO_SETOPT) {
418 1.108 plunky error = sockopt_getint(sopt, &optval);
419 1.108 plunky if (error)
420 1.108 plunky goto out;
421 1.108 plunky if (optval) {
422 1.108 plunky inp->inp_flags &= ~INP_HDRINCL;
423 1.108 plunky inp->inp_flags |= INP_NOHEADER;
424 1.108 plunky } else
425 1.108 plunky inp->inp_flags &= ~INP_NOHEADER;
426 1.108 plunky }
427 1.108 plunky goto out;
428 1.108 plunky } else if (sopt->sopt_level != IPPROTO_IP)
429 1.108 plunky return ip_ctloutput(op, so, sopt);
430 1.100 dyoung
431 1.100 dyoung switch (op) {
432 1.31 mycroft
433 1.31 mycroft case PRCO_SETOPT:
434 1.108 plunky switch (sopt->sopt_name) {
435 1.31 mycroft case IP_HDRINCL:
436 1.108 plunky error = sockopt_getint(sopt, &optval);
437 1.108 plunky if (error)
438 1.108 plunky break;
439 1.108 plunky if (optval)
440 1.100 dyoung inp->inp_flags |= INP_HDRINCL;
441 1.100 dyoung else
442 1.100 dyoung inp->inp_flags &= ~INP_HDRINCL;
443 1.108 plunky break;
444 1.31 mycroft
445 1.31 mycroft #ifdef MROUTING
446 1.31 mycroft case MRT_INIT:
447 1.31 mycroft case MRT_DONE:
448 1.31 mycroft case MRT_ADD_VIF:
449 1.31 mycroft case MRT_DEL_VIF:
450 1.31 mycroft case MRT_ADD_MFC:
451 1.31 mycroft case MRT_DEL_MFC:
452 1.31 mycroft case MRT_ASSERT:
453 1.81 manu case MRT_API_CONFIG:
454 1.81 manu case MRT_ADD_BW_UPCALL:
455 1.81 manu case MRT_DEL_BW_UPCALL:
456 1.108 plunky error = ip_mrouter_set(so, sopt);
457 1.31 mycroft break;
458 1.31 mycroft #endif
459 1.31 mycroft
460 1.31 mycroft default:
461 1.108 plunky error = ip_ctloutput(op, so, sopt);
462 1.31 mycroft break;
463 1.13 mycroft }
464 1.13 mycroft break;
465 1.1 cgd
466 1.31 mycroft case PRCO_GETOPT:
467 1.108 plunky switch (sopt->sopt_name) {
468 1.31 mycroft case IP_HDRINCL:
469 1.108 plunky optval = inp->inp_flags & INP_HDRINCL;
470 1.108 plunky error = sockopt_set(sopt, &optval, sizeof(optval));
471 1.31 mycroft break;
472 1.31 mycroft
473 1.6 hpeyerl #ifdef MROUTING
474 1.31 mycroft case MRT_VERSION:
475 1.31 mycroft case MRT_ASSERT:
476 1.81 manu case MRT_API_SUPPORT:
477 1.81 manu case MRT_API_CONFIG:
478 1.108 plunky error = ip_mrouter_get(so, sopt);
479 1.18 mycroft break;
480 1.31 mycroft #endif
481 1.31 mycroft
482 1.18 mycroft default:
483 1.108 plunky error = ip_ctloutput(op, so, sopt);
484 1.18 mycroft break;
485 1.18 mycroft }
486 1.31 mycroft break;
487 1.1 cgd }
488 1.108 plunky out:
489 1.100 dyoung return error;
490 1.1 cgd }
491 1.1 cgd
492 1.27 mycroft int
493 1.149 rtr rip_connect_pcb(struct inpcb *inp, struct sockaddr_in *addr)
494 1.27 mycroft {
495 1.27 mycroft
496 1.158 ozaki if (IFNET_READER_EMPTY())
497 1.27 mycroft return (EADDRNOTAVAIL);
498 1.115 joerg if (addr->sin_family != AF_INET)
499 1.27 mycroft return (EAFNOSUPPORT);
500 1.27 mycroft inp->inp_faddr = addr->sin_addr;
501 1.27 mycroft return (0);
502 1.27 mycroft }
503 1.27 mycroft
504 1.139 rtr static void
505 1.139 rtr rip_disconnect1(struct inpcb *inp)
506 1.27 mycroft {
507 1.27 mycroft
508 1.32 mycroft inp->inp_faddr = zeroin_addr;
509 1.27 mycroft }
510 1.27 mycroft
511 1.121 rmind static int
512 1.121 rmind rip_attach(struct socket *so, int proto)
513 1.121 rmind {
514 1.121 rmind struct inpcb *inp;
515 1.121 rmind int error;
516 1.121 rmind
517 1.121 rmind KASSERT(sotoinpcb(so) == NULL);
518 1.121 rmind sosetlock(so);
519 1.121 rmind
520 1.121 rmind if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
521 1.121 rmind error = soreserve(so, rip_sendspace, rip_recvspace);
522 1.121 rmind if (error) {
523 1.121 rmind return error;
524 1.121 rmind }
525 1.121 rmind }
526 1.121 rmind
527 1.121 rmind error = in_pcballoc(so, &rawcbtable);
528 1.121 rmind if (error) {
529 1.121 rmind return error;
530 1.121 rmind }
531 1.121 rmind inp = sotoinpcb(so);
532 1.121 rmind inp->inp_ip.ip_p = proto;
533 1.121 rmind KASSERT(solocked(so));
534 1.121 rmind
535 1.121 rmind return 0;
536 1.121 rmind }
537 1.13 mycroft
538 1.121 rmind static void
539 1.121 rmind rip_detach(struct socket *so)
540 1.1 cgd {
541 1.53 augustss struct inpcb *inp;
542 1.121 rmind
543 1.121 rmind KASSERT(solocked(so));
544 1.121 rmind inp = sotoinpcb(so);
545 1.121 rmind KASSERT(inp != NULL);
546 1.121 rmind
547 1.13 mycroft #ifdef MROUTING
548 1.6 hpeyerl extern struct socket *ip_mrouter;
549 1.121 rmind if (so == ip_mrouter) {
550 1.121 rmind ip_mrouter_done();
551 1.121 rmind }
552 1.6 hpeyerl #endif
553 1.121 rmind in_pcbdetach(inp);
554 1.121 rmind }
555 1.121 rmind
556 1.125 rtr static int
557 1.148 rtr rip_accept(struct socket *so, struct sockaddr *nam)
558 1.133 rtr {
559 1.133 rtr KASSERT(solocked(so));
560 1.133 rtr
561 1.133 rtr panic("rip_accept");
562 1.136 rtr
563 1.136 rtr return EOPNOTSUPP;
564 1.136 rtr }
565 1.136 rtr
566 1.136 rtr static int
567 1.147 rtr rip_bind(struct socket *so, struct sockaddr *nam, struct lwp *l)
568 1.136 rtr {
569 1.136 rtr struct inpcb *inp = sotoinpcb(so);
570 1.147 rtr struct sockaddr_in *addr = (struct sockaddr_in *)nam;
571 1.136 rtr int error = 0;
572 1.159 ozaki int s, ss;
573 1.159 ozaki struct ifaddr *ifa;
574 1.136 rtr
575 1.136 rtr KASSERT(solocked(so));
576 1.136 rtr KASSERT(inp != NULL);
577 1.136 rtr KASSERT(nam != NULL);
578 1.136 rtr
579 1.147 rtr if (addr->sin_len != sizeof(*addr))
580 1.147 rtr return EINVAL;
581 1.147 rtr
582 1.136 rtr s = splsoftnet();
583 1.158 ozaki if (IFNET_READER_EMPTY()) {
584 1.136 rtr error = EADDRNOTAVAIL;
585 1.136 rtr goto release;
586 1.136 rtr }
587 1.136 rtr if (addr->sin_family != AF_INET) {
588 1.136 rtr error = EAFNOSUPPORT;
589 1.136 rtr goto release;
590 1.136 rtr }
591 1.159 ozaki ss = pserialize_read_enter();
592 1.159 ozaki if ((ifa = ifa_ifwithaddr(sintosa(addr))) == NULL &&
593 1.151 roy !in_nullhost(addr->sin_addr))
594 1.151 roy {
595 1.159 ozaki pserialize_read_exit(ss);
596 1.136 rtr error = EADDRNOTAVAIL;
597 1.136 rtr goto release;
598 1.136 rtr }
599 1.161 roy if (ifa && (ifatoia(ifa))->ia4_flags & IN6_IFF_DUPLICATED) {
600 1.159 ozaki pserialize_read_exit(ss);
601 1.151 roy error = EADDRNOTAVAIL;
602 1.151 roy goto release;
603 1.151 roy }
604 1.159 ozaki pserialize_read_exit(ss);
605 1.151 roy
606 1.136 rtr inp->inp_laddr = addr->sin_addr;
607 1.136 rtr
608 1.136 rtr release:
609 1.136 rtr splx(s);
610 1.136 rtr return error;
611 1.136 rtr }
612 1.136 rtr
613 1.136 rtr static int
614 1.142 rtr rip_listen(struct socket *so, struct lwp *l)
615 1.136 rtr {
616 1.136 rtr KASSERT(solocked(so));
617 1.136 rtr
618 1.133 rtr return EOPNOTSUPP;
619 1.133 rtr }
620 1.133 rtr
621 1.133 rtr static int
622 1.152 rtr rip_connect(struct socket *so, struct sockaddr *nam, struct lwp *l)
623 1.137 rtr {
624 1.137 rtr struct inpcb *inp = sotoinpcb(so);
625 1.137 rtr int error = 0;
626 1.137 rtr int s;
627 1.137 rtr
628 1.137 rtr KASSERT(solocked(so));
629 1.137 rtr KASSERT(inp != NULL);
630 1.137 rtr KASSERT(nam != NULL);
631 1.137 rtr
632 1.137 rtr s = splsoftnet();
633 1.152 rtr error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
634 1.137 rtr if (! error)
635 1.137 rtr soisconnected(so);
636 1.140 rtr splx(s);
637 1.137 rtr
638 1.137 rtr return error;
639 1.137 rtr }
640 1.137 rtr
641 1.139 rtr static int
642 1.145 rtr rip_connect2(struct socket *so, struct socket *so2)
643 1.145 rtr {
644 1.145 rtr KASSERT(solocked(so));
645 1.145 rtr
646 1.145 rtr return EOPNOTSUPP;
647 1.145 rtr }
648 1.145 rtr
649 1.145 rtr static int
650 1.139 rtr rip_disconnect(struct socket *so)
651 1.139 rtr {
652 1.139 rtr struct inpcb *inp = sotoinpcb(so);
653 1.140 rtr int s;
654 1.139 rtr
655 1.139 rtr KASSERT(solocked(so));
656 1.139 rtr KASSERT(inp != NULL);
657 1.139 rtr
658 1.140 rtr s = splsoftnet();
659 1.139 rtr soisdisconnected(so);
660 1.139 rtr rip_disconnect1(inp);
661 1.140 rtr splx(s);
662 1.140 rtr
663 1.139 rtr return 0;
664 1.139 rtr }
665 1.139 rtr
666 1.139 rtr static int
667 1.139 rtr rip_shutdown(struct socket *so)
668 1.139 rtr {
669 1.140 rtr int s;
670 1.140 rtr
671 1.139 rtr KASSERT(solocked(so));
672 1.139 rtr
673 1.139 rtr /*
674 1.139 rtr * Mark the connection as being incapable of further input.
675 1.139 rtr */
676 1.140 rtr s = splsoftnet();
677 1.139 rtr socantsendmore(so);
678 1.140 rtr splx(s);
679 1.140 rtr
680 1.139 rtr return 0;
681 1.139 rtr }
682 1.139 rtr
683 1.139 rtr static int
684 1.139 rtr rip_abort(struct socket *so)
685 1.139 rtr {
686 1.139 rtr KASSERT(solocked(so));
687 1.139 rtr
688 1.139 rtr panic("rip_abort");
689 1.139 rtr
690 1.139 rtr return EOPNOTSUPP;
691 1.139 rtr }
692 1.137 rtr
693 1.137 rtr static int
694 1.127 rtr rip_ioctl(struct socket *so, u_long cmd, void *nam, struct ifnet *ifp)
695 1.125 rtr {
696 1.127 rtr return in_control(so, cmd, nam, ifp);
697 1.125 rtr }
698 1.125 rtr
699 1.128 rtr static int
700 1.128 rtr rip_stat(struct socket *so, struct stat *ub)
701 1.128 rtr {
702 1.131 rtr KASSERT(solocked(so));
703 1.131 rtr
704 1.130 rtr /* stat: don't bother with a blocksize. */
705 1.130 rtr return 0;
706 1.128 rtr }
707 1.128 rtr
708 1.132 rtr static int
709 1.148 rtr rip_peeraddr(struct socket *so, struct sockaddr *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 1.148 rtr in_setpeeraddr(sotoinpcb(so), (struct sockaddr_in *)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 1.132 rtr static int
725 1.148 rtr rip_sockaddr(struct socket *so, struct sockaddr *nam)
726 1.132 rtr {
727 1.140 rtr int s;
728 1.140 rtr
729 1.134 rtr KASSERT(solocked(so));
730 1.132 rtr KASSERT(sotoinpcb(so) != NULL);
731 1.132 rtr KASSERT(nam != NULL);
732 1.132 rtr
733 1.140 rtr s = splsoftnet();
734 1.148 rtr in_setsockaddr(sotoinpcb(so), (struct sockaddr_in *)nam);
735 1.140 rtr splx(s);
736 1.140 rtr
737 1.132 rtr return 0;
738 1.132 rtr }
739 1.132 rtr
740 1.135 rtr static int
741 1.144 rtr rip_rcvd(struct socket *so, int flags, struct lwp *l)
742 1.144 rtr {
743 1.144 rtr KASSERT(solocked(so));
744 1.144 rtr
745 1.144 rtr return EOPNOTSUPP;
746 1.144 rtr }
747 1.144 rtr
748 1.144 rtr static int
749 1.135 rtr rip_recvoob(struct socket *so, struct mbuf *m, int flags)
750 1.135 rtr {
751 1.135 rtr KASSERT(solocked(so));
752 1.135 rtr
753 1.135 rtr return EOPNOTSUPP;
754 1.135 rtr }
755 1.135 rtr
756 1.135 rtr static int
757 1.152 rtr rip_send(struct socket *so, struct mbuf *m, struct sockaddr *nam,
758 1.143 rtr struct mbuf *control, struct lwp *l)
759 1.143 rtr {
760 1.143 rtr struct inpcb *inp = sotoinpcb(so);
761 1.143 rtr int error = 0;
762 1.143 rtr int s;
763 1.143 rtr
764 1.143 rtr KASSERT(solocked(so));
765 1.143 rtr KASSERT(inp != NULL);
766 1.143 rtr KASSERT(m != NULL);
767 1.143 rtr
768 1.143 rtr /*
769 1.143 rtr * Ship a packet out. The appropriate raw output
770 1.143 rtr * routine handles any massaging necessary.
771 1.143 rtr */
772 1.143 rtr s = splsoftnet();
773 1.143 rtr if (nam) {
774 1.143 rtr if ((so->so_state & SS_ISCONNECTED) != 0) {
775 1.143 rtr error = EISCONN;
776 1.143 rtr goto die;
777 1.143 rtr }
778 1.152 rtr error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
779 1.166 ryo if (error)
780 1.166 ryo goto die;
781 1.143 rtr } else {
782 1.143 rtr if ((so->so_state & SS_ISCONNECTED) == 0) {
783 1.143 rtr error = ENOTCONN;
784 1.143 rtr goto die;
785 1.143 rtr }
786 1.143 rtr }
787 1.166 ryo error = rip_output(m, inp, control, l);
788 1.166 ryo m = NULL;
789 1.166 ryo control = NULL;
790 1.143 rtr if (nam)
791 1.143 rtr rip_disconnect1(inp);
792 1.166 ryo die:
793 1.166 ryo if (m != NULL)
794 1.166 ryo m_freem(m);
795 1.166 ryo if (control != NULL)
796 1.166 ryo m_freem(control);
797 1.143 rtr
798 1.143 rtr splx(s);
799 1.143 rtr return error;
800 1.143 rtr }
801 1.143 rtr
802 1.143 rtr static int
803 1.135 rtr rip_sendoob(struct socket *so, struct mbuf *m, struct mbuf *control)
804 1.135 rtr {
805 1.135 rtr KASSERT(solocked(so));
806 1.135 rtr
807 1.135 rtr m_freem(m);
808 1.135 rtr m_freem(control);
809 1.135 rtr
810 1.135 rtr return EOPNOTSUPP;
811 1.135 rtr }
812 1.135 rtr
813 1.145 rtr static int
814 1.145 rtr rip_purgeif(struct socket *so, struct ifnet *ifp)
815 1.145 rtr {
816 1.145 rtr int s;
817 1.145 rtr
818 1.145 rtr s = splsoftnet();
819 1.145 rtr mutex_enter(softnet_lock);
820 1.145 rtr in_pcbpurgeif0(&rawcbtable, ifp);
821 1.162 ozaki #ifdef NET_MPSAFE
822 1.162 ozaki mutex_exit(softnet_lock);
823 1.162 ozaki #endif
824 1.145 rtr in_purgeif(ifp);
825 1.162 ozaki #ifdef NET_MPSAFE
826 1.162 ozaki mutex_enter(softnet_lock);
827 1.162 ozaki #endif
828 1.145 rtr in_pcbpurgeif(&rawcbtable, ifp);
829 1.145 rtr mutex_exit(softnet_lock);
830 1.145 rtr splx(s);
831 1.145 rtr
832 1.145 rtr return 0;
833 1.145 rtr }
834 1.145 rtr
835 1.122 rmind PR_WRAP_USRREQS(rip)
836 1.122 rmind #define rip_attach rip_attach_wrapper
837 1.122 rmind #define rip_detach rip_detach_wrapper
838 1.133 rtr #define rip_accept rip_accept_wrapper
839 1.136 rtr #define rip_bind rip_bind_wrapper
840 1.136 rtr #define rip_listen rip_listen_wrapper
841 1.137 rtr #define rip_connect rip_connect_wrapper
842 1.145 rtr #define rip_connect2 rip_connect2_wrapper
843 1.139 rtr #define rip_disconnect rip_disconnect_wrapper
844 1.139 rtr #define rip_shutdown rip_shutdown_wrapper
845 1.139 rtr #define rip_abort rip_abort_wrapper
846 1.125 rtr #define rip_ioctl rip_ioctl_wrapper
847 1.128 rtr #define rip_stat rip_stat_wrapper
848 1.132 rtr #define rip_peeraddr rip_peeraddr_wrapper
849 1.132 rtr #define rip_sockaddr rip_sockaddr_wrapper
850 1.144 rtr #define rip_rcvd rip_rcvd_wrapper
851 1.135 rtr #define rip_recvoob rip_recvoob_wrapper
852 1.143 rtr #define rip_send rip_send_wrapper
853 1.135 rtr #define rip_sendoob rip_sendoob_wrapper
854 1.145 rtr #define rip_purgeif rip_purgeif_wrapper
855 1.120 rmind
856 1.120 rmind const struct pr_usrreqs rip_usrreqs = {
857 1.121 rmind .pr_attach = rip_attach,
858 1.121 rmind .pr_detach = rip_detach,
859 1.133 rtr .pr_accept = rip_accept,
860 1.136 rtr .pr_bind = rip_bind,
861 1.136 rtr .pr_listen = rip_listen,
862 1.137 rtr .pr_connect = rip_connect,
863 1.145 rtr .pr_connect2 = rip_connect2,
864 1.139 rtr .pr_disconnect = rip_disconnect,
865 1.139 rtr .pr_shutdown = rip_shutdown,
866 1.139 rtr .pr_abort = rip_abort,
867 1.125 rtr .pr_ioctl = rip_ioctl,
868 1.128 rtr .pr_stat = rip_stat,
869 1.132 rtr .pr_peeraddr = rip_peeraddr,
870 1.132 rtr .pr_sockaddr = rip_sockaddr,
871 1.144 rtr .pr_rcvd = rip_rcvd,
872 1.135 rtr .pr_recvoob = rip_recvoob,
873 1.143 rtr .pr_send = rip_send,
874 1.135 rtr .pr_sendoob = rip_sendoob,
875 1.145 rtr .pr_purgeif = rip_purgeif,
876 1.120 rmind };
877 1.120 rmind
878 1.110 pooka static void
879 1.110 pooka sysctl_net_inet_raw_setup(struct sysctllog **clog)
880 1.84 atatat {
881 1.84 atatat
882 1.84 atatat sysctl_createv(clog, 0, NULL, NULL,
883 1.84 atatat CTLFLAG_PERMANENT,
884 1.84 atatat CTLTYPE_NODE, "inet", NULL,
885 1.84 atatat NULL, 0, NULL, 0,
886 1.84 atatat CTL_NET, PF_INET, CTL_EOL);
887 1.84 atatat sysctl_createv(clog, 0, NULL, NULL,
888 1.84 atatat CTLFLAG_PERMANENT,
889 1.84 atatat CTLTYPE_NODE, "raw",
890 1.84 atatat SYSCTL_DESCR("Raw IPv4 settings"),
891 1.84 atatat NULL, 0, NULL, 0,
892 1.84 atatat CTL_NET, PF_INET, IPPROTO_RAW, CTL_EOL);
893 1.84 atatat
894 1.84 atatat sysctl_createv(clog, 0, NULL, NULL,
895 1.84 atatat CTLFLAG_PERMANENT,
896 1.86 atatat CTLTYPE_STRUCT, "pcblist",
897 1.84 atatat SYSCTL_DESCR("Raw IPv4 control block list"),
898 1.84 atatat sysctl_inpcblist, 0, &rawcbtable, 0,
899 1.84 atatat CTL_NET, PF_INET, IPPROTO_RAW,
900 1.84 atatat CTL_CREATE, CTL_EOL);
901 1.84 atatat }
902