raw_ip.c revision 1.172 1 1.172 roy /* $NetBSD: raw_ip.c,v 1.172 2018/03/21 14:23:54 roy 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 1.172 roy __KERNEL_RCSID(0, "$NetBSD: raw_ip.c,v 1.172 2018/03/21 14:23:54 roy 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 1.171 maxv #endif
105 1.72 jonathan
106 1.20 mycroft struct inpcbtable rawcbtable;
107 1.30 pk
108 1.82 perry int rip_pcbnotify(struct inpcbtable *, struct in_addr,
109 1.82 perry struct in_addr, int, int, void (*)(struct inpcb *, int));
110 1.149 rtr static int rip_connect_pcb(struct inpcb *, struct sockaddr_in *);
111 1.139 rtr static void rip_disconnect1(struct inpcb *);
112 1.13 mycroft
113 1.110 pooka static void sysctl_net_inet_raw_setup(struct sysctllog **);
114 1.110 pooka
115 1.13 mycroft /*
116 1.13 mycroft * Nominal space allocated to a raw ip socket.
117 1.13 mycroft */
118 1.13 mycroft #define RIPSNDQ 8192
119 1.13 mycroft #define RIPRCVQ 8192
120 1.1 cgd
121 1.121 rmind static u_long rip_sendspace = RIPSNDQ;
122 1.121 rmind static u_long rip_recvspace = RIPRCVQ;
123 1.121 rmind
124 1.1 cgd /*
125 1.1 cgd * Raw interface to IP protocol.
126 1.1 cgd */
127 1.13 mycroft
128 1.13 mycroft /*
129 1.13 mycroft * Initialize raw connection block q.
130 1.13 mycroft */
131 1.13 mycroft void
132 1.83 perry rip_init(void)
133 1.13 mycroft {
134 1.13 mycroft
135 1.110 pooka sysctl_net_inet_raw_setup(NULL);
136 1.33 mycroft in_pcbinit(&rawcbtable, 1, 1);
137 1.13 mycroft }
138 1.13 mycroft
139 1.100 dyoung static void
140 1.100 dyoung rip_sbappendaddr(struct inpcb *last, struct ip *ip, const struct sockaddr *sa,
141 1.100 dyoung int hlen, struct mbuf *opts, struct mbuf *n)
142 1.100 dyoung {
143 1.100 dyoung if (last->inp_flags & INP_NOHEADER)
144 1.100 dyoung m_adj(n, hlen);
145 1.109 christos if (last->inp_flags & INP_CONTROLOPTS
146 1.165 christos || SOOPT_TIMESTAMP(last->inp_socket->so_options))
147 1.100 dyoung ip_savecontrol(last, &opts, ip, n);
148 1.100 dyoung if (sbappendaddr(&last->inp_socket->so_rcv, sa, n, opts) == 0) {
149 1.172 roy soroverflow(last->inp_socket);
150 1.100 dyoung m_freem(n);
151 1.100 dyoung if (opts)
152 1.100 dyoung m_freem(opts);
153 1.100 dyoung } else
154 1.100 dyoung sorwakeup(last->inp_socket);
155 1.100 dyoung }
156 1.100 dyoung
157 1.1 cgd /*
158 1.1 cgd * Setup generic address and protocol structures
159 1.1 cgd * for raw_input routine, then pass them along with
160 1.1 cgd * mbuf chain.
161 1.1 cgd */
162 1.9 mycroft void
163 1.24 christos rip_input(struct mbuf *m, ...)
164 1.1 cgd {
165 1.100 dyoung int hlen, proto;
166 1.53 augustss struct ip *ip = mtod(m, struct ip *);
167 1.75 itojun struct inpcb_hdr *inph;
168 1.53 augustss struct inpcb *inp;
169 1.97 dyoung struct inpcb *last = NULL;
170 1.97 dyoung struct mbuf *n, *opts = NULL;
171 1.32 mycroft struct sockaddr_in ripsrc;
172 1.43 itojun va_list ap;
173 1.43 itojun
174 1.43 itojun va_start(ap, m);
175 1.64 simonb (void)va_arg(ap, int); /* ignore value, advance ap */
176 1.43 itojun proto = va_arg(ap, int);
177 1.43 itojun va_end(ap);
178 1.1 cgd
179 1.97 dyoung sockaddr_in_init(&ripsrc, &ip->ip_src, 0);
180 1.42 thorpej
181 1.42 thorpej /*
182 1.42 thorpej * XXX Compatibility: programs using raw IP expect ip_len
183 1.62 itojun * XXX to have the header length subtracted, and in host order.
184 1.62 itojun * XXX ip_off is also expected to be host order.
185 1.42 thorpej */
186 1.100 dyoung hlen = ip->ip_hl << 2;
187 1.100 dyoung ip->ip_len = ntohs(ip->ip_len) - hlen;
188 1.62 itojun NTOHS(ip->ip_off);
189 1.32 mycroft
190 1.117 christos TAILQ_FOREACH(inph, &rawcbtable.inpt_queue, inph_queue) {
191 1.75 itojun inp = (struct inpcb *)inph;
192 1.75 itojun if (inp->inp_af != AF_INET)
193 1.75 itojun continue;
194 1.43 itojun if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
195 1.13 mycroft continue;
196 1.32 mycroft if (!in_nullhost(inp->inp_laddr) &&
197 1.32 mycroft !in_hosteq(inp->inp_laddr, ip->ip_dst))
198 1.13 mycroft continue;
199 1.32 mycroft if (!in_nullhost(inp->inp_faddr) &&
200 1.32 mycroft !in_hosteq(inp->inp_faddr, ip->ip_src))
201 1.13 mycroft continue;
202 1.97 dyoung if (last == NULL)
203 1.97 dyoung ;
204 1.116 christos #if defined(IPSEC)
205 1.97 dyoung /* check AH/ESP integrity. */
206 1.169 maxv else if (ipsec_used && ipsec_in_reject(m, last)) {
207 1.97 dyoung /* do not inject data to pcb */
208 1.97 dyoung }
209 1.97 dyoung #endif /*IPSEC*/
210 1.99 dyoung else if ((n = m_copypacket(m, M_DONTWAIT)) != NULL) {
211 1.100 dyoung rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts,
212 1.100 dyoung n);
213 1.97 dyoung opts = NULL;
214 1.13 mycroft }
215 1.36 thorpej last = inp;
216 1.13 mycroft }
217 1.116 christos #if defined(IPSEC)
218 1.55 itojun /* check AH/ESP integrity. */
219 1.169 maxv if (ipsec_used && last != NULL && ipsec_in_reject(m, last)) {
220 1.55 itojun m_freem(m);
221 1.105 thorpej IP_STATDEC(IP_STAT_DELIVERED);
222 1.55 itojun /* do not inject data to pcb */
223 1.55 itojun } else
224 1.55 itojun #endif /*IPSEC*/
225 1.100 dyoung if (last != NULL)
226 1.100 dyoung rip_sbappendaddr(last, ip, sintosa(&ripsrc), hlen, opts, m);
227 1.156 riastrad else if (inetsw[ip_protox[ip->ip_p]].pr_input == rip_input) {
228 1.105 thorpej uint64_t *ips;
229 1.105 thorpej
230 1.97 dyoung icmp_error(m, ICMP_UNREACH, ICMP_UNREACH_PROTOCOL,
231 1.97 dyoung 0, 0);
232 1.105 thorpej ips = IP_STAT_GETREF();
233 1.105 thorpej ips[IP_STAT_NOPROTO]++;
234 1.105 thorpej ips[IP_STAT_DELIVERED]--;
235 1.105 thorpej IP_STAT_PUTREF();
236 1.97 dyoung } else
237 1.97 dyoung m_freem(m);
238 1.43 itojun return;
239 1.60 itojun }
240 1.60 itojun
241 1.60 itojun int
242 1.83 perry rip_pcbnotify(struct inpcbtable *table,
243 1.83 perry struct in_addr faddr, struct in_addr laddr, int proto, int errno,
244 1.83 perry void (*notify)(struct inpcb *, int))
245 1.60 itojun {
246 1.117 christos struct inpcb_hdr *inph, *ninph;
247 1.60 itojun int nmatch;
248 1.60 itojun
249 1.60 itojun nmatch = 0;
250 1.117 christos TAILQ_FOREACH_SAFE(inph, &table->inpt_queue, inph_queue, ninph) {
251 1.117 christos struct inpcb *inp = (struct inpcb *)inph;
252 1.75 itojun if (inp->inp_af != AF_INET)
253 1.75 itojun continue;
254 1.60 itojun if (inp->inp_ip.ip_p && inp->inp_ip.ip_p != proto)
255 1.60 itojun continue;
256 1.60 itojun if (in_hosteq(inp->inp_faddr, faddr) &&
257 1.60 itojun in_hosteq(inp->inp_laddr, laddr)) {
258 1.60 itojun (*notify)(inp, errno);
259 1.60 itojun nmatch++;
260 1.60 itojun }
261 1.60 itojun }
262 1.60 itojun
263 1.60 itojun return nmatch;
264 1.60 itojun }
265 1.60 itojun
266 1.60 itojun void *
267 1.95 dyoung rip_ctlinput(int cmd, const struct sockaddr *sa, void *v)
268 1.60 itojun {
269 1.60 itojun struct ip *ip = v;
270 1.82 perry void (*notify)(struct inpcb *, int) = in_rtchange;
271 1.60 itojun int errno;
272 1.60 itojun
273 1.60 itojun if (sa->sa_family != AF_INET ||
274 1.60 itojun sa->sa_len != sizeof(struct sockaddr_in))
275 1.60 itojun return NULL;
276 1.60 itojun if ((unsigned)cmd >= PRC_NCMDS)
277 1.60 itojun return NULL;
278 1.60 itojun errno = inetctlerrmap[cmd];
279 1.60 itojun if (PRC_IS_REDIRECT(cmd))
280 1.60 itojun notify = in_rtchange, ip = 0;
281 1.60 itojun else if (cmd == PRC_HOSTDEAD)
282 1.60 itojun ip = 0;
283 1.60 itojun else if (errno == 0)
284 1.60 itojun return NULL;
285 1.60 itojun if (ip) {
286 1.95 dyoung rip_pcbnotify(&rawcbtable, satocsin(sa)->sin_addr,
287 1.60 itojun ip->ip_src, ip->ip_p, errno, notify);
288 1.60 itojun
289 1.60 itojun /* XXX mapped address case */
290 1.60 itojun } else
291 1.95 dyoung in_pcbnotifyall(&rawcbtable, satocsin(sa)->sin_addr, errno,
292 1.60 itojun notify);
293 1.60 itojun return NULL;
294 1.1 cgd }
295 1.1 cgd
296 1.1 cgd /*
297 1.1 cgd * Generate IP header and pass packet to ip_output.
298 1.1 cgd * Tack on options user may have setup with control call.
299 1.1 cgd */
300 1.9 mycroft int
301 1.166 ryo rip_output(struct mbuf *m, struct inpcb *inp, struct mbuf *control,
302 1.166 ryo struct lwp *l)
303 1.24 christos {
304 1.53 augustss struct ip *ip;
305 1.10 mycroft struct mbuf *opts;
306 1.166 ryo struct ip_pktopts pktopts;
307 1.166 ryo kauth_cred_t cred;
308 1.166 ryo int error, flags;
309 1.166 ryo
310 1.166 ryo flags = (inp->inp_socket->so_options & SO_DONTROUTE) |
311 1.166 ryo IP_ALLOWBROADCAST | IP_RETURNMTU;
312 1.166 ryo
313 1.166 ryo if (l == NULL)
314 1.166 ryo cred = NULL;
315 1.166 ryo else
316 1.166 ryo cred = l->l_cred;
317 1.166 ryo
318 1.166 ryo /* Setup IP outgoing packet options */
319 1.166 ryo memset(&pktopts, 0, sizeof(pktopts));
320 1.167 ryo error = ip_setpktopts(control, &pktopts, &flags, inp, cred);
321 1.166 ryo if (control != NULL)
322 1.166 ryo m_freem(control);
323 1.166 ryo if (error != 0)
324 1.166 ryo goto release;
325 1.1 cgd
326 1.1 cgd /*
327 1.1 cgd * If the user handed us a complete IP packet, use it.
328 1.1 cgd * Otherwise, allocate an mbuf for a header and fill it in.
329 1.1 cgd */
330 1.13 mycroft if ((inp->inp_flags & INP_HDRINCL) == 0) {
331 1.35 thorpej if ((m->m_pkthdr.len + sizeof(struct ip)) > IP_MAXPACKET) {
332 1.166 ryo error = EMSGSIZE;
333 1.166 ryo goto release;
334 1.35 thorpej }
335 1.68 itojun M_PREPEND(m, sizeof(struct ip), M_DONTWAIT);
336 1.166 ryo if (!m) {
337 1.166 ryo error = ENOBUFS;
338 1.166 ryo goto release;
339 1.166 ryo }
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 1.166 ryo ip->ip_src = pktopts.ippo_laddr.sin_addr;
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 1.166 ryo error = EMSGSIZE;
352 1.166 ryo goto release;
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 1.166 ryo if (m == NULL) {
366 1.166 ryo error = ENOMEM; /* XXX */
367 1.166 ryo goto release;
368 1.166 ryo }
369 1.65 thorpej ip = mtod(m, struct ip *);
370 1.65 thorpej }
371 1.65 thorpej
372 1.62 itojun /* XXX userland passes ip_len and ip_off in host order */
373 1.38 mycroft if (m->m_pkthdr.len != ip->ip_len) {
374 1.166 ryo error = EINVAL;
375 1.166 ryo goto release;
376 1.38 mycroft }
377 1.62 itojun HTONS(ip->ip_len);
378 1.62 itojun HTONS(ip->ip_off);
379 1.103 matt if (ip->ip_id != 0 || m->m_pkthdr.len < IP_MINFRAGSIZE)
380 1.103 matt flags |= IP_NOIPNEWID;
381 1.13 mycroft opts = NULL;
382 1.13 mycroft /* XXX prevent ip_output from overwriting header fields */
383 1.13 mycroft flags |= IP_RAWOUTPUT;
384 1.105 thorpej IP_STATINC(IP_STAT_RAWOUT);
385 1.1 cgd }
386 1.123 rmind
387 1.123 rmind /*
388 1.123 rmind * IP output. Note: if IP_RETURNMTU flag is set, the MTU size
389 1.123 rmind * will be stored in inp_errormtu.
390 1.123 rmind */
391 1.166 ryo return ip_output(m, opts, &inp->inp_route, flags, pktopts.ippo_imo,
392 1.166 ryo inp);
393 1.166 ryo
394 1.166 ryo release:
395 1.166 ryo if (m != NULL)
396 1.166 ryo m_freem(m);
397 1.166 ryo return error;
398 1.1 cgd }
399 1.1 cgd
400 1.1 cgd /*
401 1.1 cgd * Raw IP socket option processing.
402 1.1 cgd */
403 1.9 mycroft int
404 1.108 plunky rip_ctloutput(int op, struct socket *so, struct sockopt *sopt)
405 1.1 cgd {
406 1.53 augustss struct inpcb *inp = sotoinpcb(so);
407 1.31 mycroft int error = 0;
408 1.108 plunky int optval;
409 1.1 cgd
410 1.108 plunky if (sopt->sopt_level == SOL_SOCKET && sopt->sopt_name == SO_NOHEADER) {
411 1.100 dyoung if (op == PRCO_GETOPT) {
412 1.108 plunky optval = (inp->inp_flags & INP_NOHEADER) ? 1 : 0;
413 1.108 plunky error = sockopt_set(sopt, &optval, sizeof(optval));
414 1.108 plunky } else if (op == PRCO_SETOPT) {
415 1.108 plunky error = sockopt_getint(sopt, &optval);
416 1.108 plunky if (error)
417 1.108 plunky goto out;
418 1.108 plunky if (optval) {
419 1.108 plunky inp->inp_flags &= ~INP_HDRINCL;
420 1.108 plunky inp->inp_flags |= INP_NOHEADER;
421 1.108 plunky } else
422 1.108 plunky inp->inp_flags &= ~INP_NOHEADER;
423 1.108 plunky }
424 1.108 plunky goto out;
425 1.108 plunky } else if (sopt->sopt_level != IPPROTO_IP)
426 1.108 plunky return ip_ctloutput(op, so, sopt);
427 1.100 dyoung
428 1.100 dyoung switch (op) {
429 1.31 mycroft
430 1.31 mycroft case PRCO_SETOPT:
431 1.108 plunky switch (sopt->sopt_name) {
432 1.31 mycroft case IP_HDRINCL:
433 1.108 plunky error = sockopt_getint(sopt, &optval);
434 1.108 plunky if (error)
435 1.108 plunky break;
436 1.108 plunky if (optval)
437 1.100 dyoung inp->inp_flags |= INP_HDRINCL;
438 1.100 dyoung else
439 1.100 dyoung inp->inp_flags &= ~INP_HDRINCL;
440 1.108 plunky break;
441 1.31 mycroft
442 1.31 mycroft #ifdef MROUTING
443 1.31 mycroft case MRT_INIT:
444 1.31 mycroft case MRT_DONE:
445 1.31 mycroft case MRT_ADD_VIF:
446 1.31 mycroft case MRT_DEL_VIF:
447 1.31 mycroft case MRT_ADD_MFC:
448 1.31 mycroft case MRT_DEL_MFC:
449 1.31 mycroft case MRT_ASSERT:
450 1.81 manu case MRT_API_CONFIG:
451 1.81 manu case MRT_ADD_BW_UPCALL:
452 1.81 manu case MRT_DEL_BW_UPCALL:
453 1.108 plunky error = ip_mrouter_set(so, sopt);
454 1.31 mycroft break;
455 1.31 mycroft #endif
456 1.31 mycroft
457 1.31 mycroft default:
458 1.108 plunky error = ip_ctloutput(op, so, sopt);
459 1.31 mycroft break;
460 1.13 mycroft }
461 1.13 mycroft break;
462 1.1 cgd
463 1.31 mycroft case PRCO_GETOPT:
464 1.108 plunky switch (sopt->sopt_name) {
465 1.31 mycroft case IP_HDRINCL:
466 1.108 plunky optval = inp->inp_flags & INP_HDRINCL;
467 1.108 plunky error = sockopt_set(sopt, &optval, sizeof(optval));
468 1.31 mycroft break;
469 1.31 mycroft
470 1.6 hpeyerl #ifdef MROUTING
471 1.31 mycroft case MRT_VERSION:
472 1.31 mycroft case MRT_ASSERT:
473 1.81 manu case MRT_API_SUPPORT:
474 1.81 manu case MRT_API_CONFIG:
475 1.108 plunky error = ip_mrouter_get(so, sopt);
476 1.18 mycroft break;
477 1.31 mycroft #endif
478 1.31 mycroft
479 1.18 mycroft default:
480 1.108 plunky error = ip_ctloutput(op, so, sopt);
481 1.18 mycroft break;
482 1.18 mycroft }
483 1.31 mycroft break;
484 1.1 cgd }
485 1.108 plunky out:
486 1.100 dyoung return error;
487 1.1 cgd }
488 1.1 cgd
489 1.27 mycroft int
490 1.149 rtr rip_connect_pcb(struct inpcb *inp, struct sockaddr_in *addr)
491 1.27 mycroft {
492 1.27 mycroft
493 1.158 ozaki if (IFNET_READER_EMPTY())
494 1.27 mycroft return (EADDRNOTAVAIL);
495 1.115 joerg if (addr->sin_family != AF_INET)
496 1.27 mycroft return (EAFNOSUPPORT);
497 1.27 mycroft inp->inp_faddr = addr->sin_addr;
498 1.27 mycroft return (0);
499 1.27 mycroft }
500 1.27 mycroft
501 1.139 rtr static void
502 1.139 rtr rip_disconnect1(struct inpcb *inp)
503 1.27 mycroft {
504 1.27 mycroft
505 1.32 mycroft inp->inp_faddr = zeroin_addr;
506 1.27 mycroft }
507 1.27 mycroft
508 1.121 rmind static int
509 1.121 rmind rip_attach(struct socket *so, int proto)
510 1.121 rmind {
511 1.121 rmind struct inpcb *inp;
512 1.121 rmind int error;
513 1.121 rmind
514 1.121 rmind KASSERT(sotoinpcb(so) == NULL);
515 1.121 rmind sosetlock(so);
516 1.121 rmind
517 1.121 rmind if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
518 1.121 rmind error = soreserve(so, rip_sendspace, rip_recvspace);
519 1.121 rmind if (error) {
520 1.121 rmind return error;
521 1.121 rmind }
522 1.121 rmind }
523 1.121 rmind
524 1.121 rmind error = in_pcballoc(so, &rawcbtable);
525 1.121 rmind if (error) {
526 1.121 rmind return error;
527 1.121 rmind }
528 1.121 rmind inp = sotoinpcb(so);
529 1.121 rmind inp->inp_ip.ip_p = proto;
530 1.121 rmind KASSERT(solocked(so));
531 1.121 rmind
532 1.121 rmind return 0;
533 1.121 rmind }
534 1.13 mycroft
535 1.121 rmind static void
536 1.121 rmind rip_detach(struct socket *so)
537 1.1 cgd {
538 1.53 augustss struct inpcb *inp;
539 1.121 rmind
540 1.121 rmind KASSERT(solocked(so));
541 1.121 rmind inp = sotoinpcb(so);
542 1.121 rmind KASSERT(inp != NULL);
543 1.121 rmind
544 1.13 mycroft #ifdef MROUTING
545 1.6 hpeyerl extern struct socket *ip_mrouter;
546 1.121 rmind if (so == ip_mrouter) {
547 1.121 rmind ip_mrouter_done();
548 1.121 rmind }
549 1.6 hpeyerl #endif
550 1.121 rmind in_pcbdetach(inp);
551 1.121 rmind }
552 1.121 rmind
553 1.125 rtr static int
554 1.148 rtr rip_accept(struct socket *so, struct sockaddr *nam)
555 1.133 rtr {
556 1.133 rtr KASSERT(solocked(so));
557 1.133 rtr
558 1.133 rtr panic("rip_accept");
559 1.136 rtr
560 1.136 rtr return EOPNOTSUPP;
561 1.136 rtr }
562 1.136 rtr
563 1.136 rtr static int
564 1.147 rtr rip_bind(struct socket *so, struct sockaddr *nam, struct lwp *l)
565 1.136 rtr {
566 1.136 rtr struct inpcb *inp = sotoinpcb(so);
567 1.147 rtr struct sockaddr_in *addr = (struct sockaddr_in *)nam;
568 1.136 rtr int error = 0;
569 1.159 ozaki int s, ss;
570 1.159 ozaki struct ifaddr *ifa;
571 1.136 rtr
572 1.136 rtr KASSERT(solocked(so));
573 1.136 rtr KASSERT(inp != NULL);
574 1.136 rtr KASSERT(nam != NULL);
575 1.136 rtr
576 1.147 rtr if (addr->sin_len != sizeof(*addr))
577 1.147 rtr return EINVAL;
578 1.147 rtr
579 1.136 rtr s = splsoftnet();
580 1.158 ozaki if (IFNET_READER_EMPTY()) {
581 1.136 rtr error = EADDRNOTAVAIL;
582 1.136 rtr goto release;
583 1.136 rtr }
584 1.136 rtr if (addr->sin_family != AF_INET) {
585 1.136 rtr error = EAFNOSUPPORT;
586 1.136 rtr goto release;
587 1.136 rtr }
588 1.159 ozaki ss = pserialize_read_enter();
589 1.159 ozaki if ((ifa = ifa_ifwithaddr(sintosa(addr))) == NULL &&
590 1.151 roy !in_nullhost(addr->sin_addr))
591 1.151 roy {
592 1.159 ozaki pserialize_read_exit(ss);
593 1.136 rtr error = EADDRNOTAVAIL;
594 1.136 rtr goto release;
595 1.136 rtr }
596 1.161 roy if (ifa && (ifatoia(ifa))->ia4_flags & IN6_IFF_DUPLICATED) {
597 1.159 ozaki pserialize_read_exit(ss);
598 1.151 roy error = EADDRNOTAVAIL;
599 1.151 roy goto release;
600 1.151 roy }
601 1.159 ozaki pserialize_read_exit(ss);
602 1.151 roy
603 1.136 rtr inp->inp_laddr = addr->sin_addr;
604 1.136 rtr
605 1.136 rtr release:
606 1.136 rtr splx(s);
607 1.136 rtr return error;
608 1.136 rtr }
609 1.136 rtr
610 1.136 rtr static int
611 1.142 rtr rip_listen(struct socket *so, struct lwp *l)
612 1.136 rtr {
613 1.136 rtr KASSERT(solocked(so));
614 1.136 rtr
615 1.133 rtr return EOPNOTSUPP;
616 1.133 rtr }
617 1.133 rtr
618 1.133 rtr static int
619 1.152 rtr rip_connect(struct socket *so, struct sockaddr *nam, struct lwp *l)
620 1.137 rtr {
621 1.137 rtr struct inpcb *inp = sotoinpcb(so);
622 1.137 rtr int error = 0;
623 1.137 rtr int s;
624 1.137 rtr
625 1.137 rtr KASSERT(solocked(so));
626 1.137 rtr KASSERT(inp != NULL);
627 1.137 rtr KASSERT(nam != NULL);
628 1.137 rtr
629 1.137 rtr s = splsoftnet();
630 1.152 rtr error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
631 1.137 rtr if (! error)
632 1.137 rtr soisconnected(so);
633 1.140 rtr splx(s);
634 1.137 rtr
635 1.137 rtr return error;
636 1.137 rtr }
637 1.137 rtr
638 1.139 rtr static int
639 1.145 rtr rip_connect2(struct socket *so, struct socket *so2)
640 1.145 rtr {
641 1.145 rtr KASSERT(solocked(so));
642 1.145 rtr
643 1.145 rtr return EOPNOTSUPP;
644 1.145 rtr }
645 1.145 rtr
646 1.145 rtr static int
647 1.139 rtr rip_disconnect(struct socket *so)
648 1.139 rtr {
649 1.139 rtr struct inpcb *inp = sotoinpcb(so);
650 1.140 rtr int s;
651 1.139 rtr
652 1.139 rtr KASSERT(solocked(so));
653 1.139 rtr KASSERT(inp != NULL);
654 1.139 rtr
655 1.140 rtr s = splsoftnet();
656 1.139 rtr soisdisconnected(so);
657 1.139 rtr rip_disconnect1(inp);
658 1.140 rtr splx(s);
659 1.140 rtr
660 1.139 rtr return 0;
661 1.139 rtr }
662 1.139 rtr
663 1.139 rtr static int
664 1.139 rtr rip_shutdown(struct socket *so)
665 1.139 rtr {
666 1.140 rtr int s;
667 1.140 rtr
668 1.139 rtr KASSERT(solocked(so));
669 1.139 rtr
670 1.139 rtr /*
671 1.139 rtr * Mark the connection as being incapable of further input.
672 1.139 rtr */
673 1.140 rtr s = splsoftnet();
674 1.139 rtr socantsendmore(so);
675 1.140 rtr splx(s);
676 1.140 rtr
677 1.139 rtr return 0;
678 1.139 rtr }
679 1.139 rtr
680 1.139 rtr static int
681 1.139 rtr rip_abort(struct socket *so)
682 1.139 rtr {
683 1.139 rtr KASSERT(solocked(so));
684 1.139 rtr
685 1.139 rtr panic("rip_abort");
686 1.139 rtr
687 1.139 rtr return EOPNOTSUPP;
688 1.139 rtr }
689 1.137 rtr
690 1.137 rtr static int
691 1.127 rtr rip_ioctl(struct socket *so, u_long cmd, void *nam, struct ifnet *ifp)
692 1.125 rtr {
693 1.127 rtr return in_control(so, cmd, nam, ifp);
694 1.125 rtr }
695 1.125 rtr
696 1.128 rtr static int
697 1.128 rtr rip_stat(struct socket *so, struct stat *ub)
698 1.128 rtr {
699 1.131 rtr KASSERT(solocked(so));
700 1.131 rtr
701 1.130 rtr /* stat: don't bother with a blocksize. */
702 1.130 rtr return 0;
703 1.128 rtr }
704 1.128 rtr
705 1.132 rtr static int
706 1.148 rtr rip_peeraddr(struct socket *so, struct sockaddr *nam)
707 1.132 rtr {
708 1.140 rtr int s;
709 1.140 rtr
710 1.134 rtr KASSERT(solocked(so));
711 1.132 rtr KASSERT(sotoinpcb(so) != NULL);
712 1.132 rtr KASSERT(nam != NULL);
713 1.132 rtr
714 1.140 rtr s = splsoftnet();
715 1.148 rtr in_setpeeraddr(sotoinpcb(so), (struct sockaddr_in *)nam);
716 1.140 rtr splx(s);
717 1.140 rtr
718 1.132 rtr return 0;
719 1.132 rtr }
720 1.132 rtr
721 1.132 rtr static int
722 1.148 rtr rip_sockaddr(struct socket *so, struct sockaddr *nam)
723 1.132 rtr {
724 1.140 rtr int s;
725 1.140 rtr
726 1.134 rtr KASSERT(solocked(so));
727 1.132 rtr KASSERT(sotoinpcb(so) != NULL);
728 1.132 rtr KASSERT(nam != NULL);
729 1.132 rtr
730 1.140 rtr s = splsoftnet();
731 1.148 rtr in_setsockaddr(sotoinpcb(so), (struct sockaddr_in *)nam);
732 1.140 rtr splx(s);
733 1.140 rtr
734 1.132 rtr return 0;
735 1.132 rtr }
736 1.132 rtr
737 1.135 rtr static int
738 1.144 rtr rip_rcvd(struct socket *so, int flags, struct lwp *l)
739 1.144 rtr {
740 1.144 rtr KASSERT(solocked(so));
741 1.144 rtr
742 1.144 rtr return EOPNOTSUPP;
743 1.144 rtr }
744 1.144 rtr
745 1.144 rtr static int
746 1.135 rtr rip_recvoob(struct socket *so, struct mbuf *m, int flags)
747 1.135 rtr {
748 1.135 rtr KASSERT(solocked(so));
749 1.135 rtr
750 1.135 rtr return EOPNOTSUPP;
751 1.135 rtr }
752 1.135 rtr
753 1.135 rtr static int
754 1.152 rtr rip_send(struct socket *so, struct mbuf *m, struct sockaddr *nam,
755 1.143 rtr struct mbuf *control, struct lwp *l)
756 1.143 rtr {
757 1.143 rtr struct inpcb *inp = sotoinpcb(so);
758 1.143 rtr int error = 0;
759 1.143 rtr int s;
760 1.143 rtr
761 1.143 rtr KASSERT(solocked(so));
762 1.143 rtr KASSERT(inp != NULL);
763 1.143 rtr KASSERT(m != NULL);
764 1.143 rtr
765 1.143 rtr /*
766 1.143 rtr * Ship a packet out. The appropriate raw output
767 1.143 rtr * routine handles any massaging necessary.
768 1.143 rtr */
769 1.143 rtr s = splsoftnet();
770 1.143 rtr if (nam) {
771 1.143 rtr if ((so->so_state & SS_ISCONNECTED) != 0) {
772 1.143 rtr error = EISCONN;
773 1.143 rtr goto die;
774 1.143 rtr }
775 1.152 rtr error = rip_connect_pcb(inp, (struct sockaddr_in *)nam);
776 1.166 ryo if (error)
777 1.166 ryo goto die;
778 1.143 rtr } else {
779 1.143 rtr if ((so->so_state & SS_ISCONNECTED) == 0) {
780 1.143 rtr error = ENOTCONN;
781 1.143 rtr goto die;
782 1.143 rtr }
783 1.143 rtr }
784 1.166 ryo error = rip_output(m, inp, control, l);
785 1.166 ryo m = NULL;
786 1.166 ryo control = NULL;
787 1.143 rtr if (nam)
788 1.143 rtr rip_disconnect1(inp);
789 1.166 ryo die:
790 1.166 ryo if (m != NULL)
791 1.166 ryo m_freem(m);
792 1.166 ryo if (control != NULL)
793 1.166 ryo m_freem(control);
794 1.143 rtr
795 1.143 rtr splx(s);
796 1.143 rtr return error;
797 1.143 rtr }
798 1.143 rtr
799 1.143 rtr static int
800 1.135 rtr rip_sendoob(struct socket *so, struct mbuf *m, struct mbuf *control)
801 1.135 rtr {
802 1.135 rtr KASSERT(solocked(so));
803 1.135 rtr
804 1.135 rtr m_freem(m);
805 1.135 rtr m_freem(control);
806 1.135 rtr
807 1.135 rtr return EOPNOTSUPP;
808 1.135 rtr }
809 1.135 rtr
810 1.145 rtr static int
811 1.145 rtr rip_purgeif(struct socket *so, struct ifnet *ifp)
812 1.145 rtr {
813 1.145 rtr int s;
814 1.145 rtr
815 1.145 rtr s = splsoftnet();
816 1.145 rtr mutex_enter(softnet_lock);
817 1.145 rtr in_pcbpurgeif0(&rawcbtable, ifp);
818 1.162 ozaki #ifdef NET_MPSAFE
819 1.162 ozaki mutex_exit(softnet_lock);
820 1.162 ozaki #endif
821 1.145 rtr in_purgeif(ifp);
822 1.162 ozaki #ifdef NET_MPSAFE
823 1.162 ozaki mutex_enter(softnet_lock);
824 1.162 ozaki #endif
825 1.145 rtr in_pcbpurgeif(&rawcbtable, ifp);
826 1.145 rtr mutex_exit(softnet_lock);
827 1.145 rtr splx(s);
828 1.145 rtr
829 1.145 rtr return 0;
830 1.145 rtr }
831 1.145 rtr
832 1.122 rmind PR_WRAP_USRREQS(rip)
833 1.122 rmind #define rip_attach rip_attach_wrapper
834 1.122 rmind #define rip_detach rip_detach_wrapper
835 1.133 rtr #define rip_accept rip_accept_wrapper
836 1.136 rtr #define rip_bind rip_bind_wrapper
837 1.136 rtr #define rip_listen rip_listen_wrapper
838 1.137 rtr #define rip_connect rip_connect_wrapper
839 1.145 rtr #define rip_connect2 rip_connect2_wrapper
840 1.139 rtr #define rip_disconnect rip_disconnect_wrapper
841 1.139 rtr #define rip_shutdown rip_shutdown_wrapper
842 1.139 rtr #define rip_abort rip_abort_wrapper
843 1.125 rtr #define rip_ioctl rip_ioctl_wrapper
844 1.128 rtr #define rip_stat rip_stat_wrapper
845 1.132 rtr #define rip_peeraddr rip_peeraddr_wrapper
846 1.132 rtr #define rip_sockaddr rip_sockaddr_wrapper
847 1.144 rtr #define rip_rcvd rip_rcvd_wrapper
848 1.135 rtr #define rip_recvoob rip_recvoob_wrapper
849 1.143 rtr #define rip_send rip_send_wrapper
850 1.135 rtr #define rip_sendoob rip_sendoob_wrapper
851 1.145 rtr #define rip_purgeif rip_purgeif_wrapper
852 1.120 rmind
853 1.120 rmind const struct pr_usrreqs rip_usrreqs = {
854 1.121 rmind .pr_attach = rip_attach,
855 1.121 rmind .pr_detach = rip_detach,
856 1.133 rtr .pr_accept = rip_accept,
857 1.136 rtr .pr_bind = rip_bind,
858 1.136 rtr .pr_listen = rip_listen,
859 1.137 rtr .pr_connect = rip_connect,
860 1.145 rtr .pr_connect2 = rip_connect2,
861 1.139 rtr .pr_disconnect = rip_disconnect,
862 1.139 rtr .pr_shutdown = rip_shutdown,
863 1.139 rtr .pr_abort = rip_abort,
864 1.125 rtr .pr_ioctl = rip_ioctl,
865 1.128 rtr .pr_stat = rip_stat,
866 1.132 rtr .pr_peeraddr = rip_peeraddr,
867 1.132 rtr .pr_sockaddr = rip_sockaddr,
868 1.144 rtr .pr_rcvd = rip_rcvd,
869 1.135 rtr .pr_recvoob = rip_recvoob,
870 1.143 rtr .pr_send = rip_send,
871 1.135 rtr .pr_sendoob = rip_sendoob,
872 1.145 rtr .pr_purgeif = rip_purgeif,
873 1.120 rmind };
874 1.120 rmind
875 1.110 pooka static void
876 1.110 pooka sysctl_net_inet_raw_setup(struct sysctllog **clog)
877 1.84 atatat {
878 1.84 atatat
879 1.84 atatat sysctl_createv(clog, 0, NULL, NULL,
880 1.84 atatat CTLFLAG_PERMANENT,
881 1.84 atatat CTLTYPE_NODE, "inet", NULL,
882 1.84 atatat NULL, 0, NULL, 0,
883 1.84 atatat CTL_NET, PF_INET, CTL_EOL);
884 1.84 atatat sysctl_createv(clog, 0, NULL, NULL,
885 1.84 atatat CTLFLAG_PERMANENT,
886 1.84 atatat CTLTYPE_NODE, "raw",
887 1.84 atatat SYSCTL_DESCR("Raw IPv4 settings"),
888 1.84 atatat NULL, 0, NULL, 0,
889 1.84 atatat CTL_NET, PF_INET, IPPROTO_RAW, CTL_EOL);
890 1.84 atatat
891 1.84 atatat sysctl_createv(clog, 0, NULL, NULL,
892 1.84 atatat CTLFLAG_PERMANENT,
893 1.86 atatat CTLTYPE_STRUCT, "pcblist",
894 1.84 atatat SYSCTL_DESCR("Raw IPv4 control block list"),
895 1.84 atatat sysctl_inpcblist, 0, &rawcbtable, 0,
896 1.84 atatat CTL_NET, PF_INET, IPPROTO_RAW,
897 1.84 atatat CTL_CREATE, CTL_EOL);
898 1.84 atatat }
899