1/*
2 * Copyright (c) 1982, 1986, 1988, 1993
3 * The Regents of the University of California. All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * 3. All advertising materials mentioning features or use of this software
14 * must display the following acknowledgement:
15 * This product includes software developed by the University of
16 * California, Berkeley and its contributors.
17 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
20 *
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
31 * SUCH DAMAGE.
32 *
33 * $Id: ip_divert.c,v 1.22 1998/03/28 10:18:23 bde Exp $
34 */
35
36#include "opt_inet.h"
37
38#ifndef INET
39#error "IPDIVERT requires INET."
40#endif
41
42#include <sys/param.h>
43#include <sys/malloc.h>
44#include <sys/mbuf.h>
45#include <sys/socket.h>
46#include <sys/protosw.h>
47#include <sys/socketvar.h>
48#include <sys/systm.h>
49#include <sys/proc.h>
50
51#include <vm/vm_zone.h>
52
53#include <net/if.h>
54#include <net/route.h>
55
56#include <netinet/in.h>
57#include <netinet/in_systm.h>
58#include <netinet/ip.h>
59#include <netinet/in_pcb.h>
60#include <netinet/in_var.h>
61#include <netinet/ip_var.h>
62
63/*
64 * Divert sockets
65 */
66
67/*
68 * Allocate enough space to hold a full IP packet
69 */
70#define DIVSNDQ (65536 + 100)
71#define DIVRCVQ (65536 + 100)
72
73/* Global variables */
74
75/*
76 * ip_input() and ip_output() set this secret value before calling us to
77 * let us know which divert port to divert a packet to; this is done so
78 * we can use the existing prototype for struct protosw's pr_input().
79 * This is stored in host order.
80 */
81u_short ip_divert_port;
82
83/*
84 * We set this value to a non-zero port number when we want the call to
85 * ip_fw_chk() in ip_input() or ip_output() to ignore ``divert <port>''
86 * chain entries. This is stored in host order.
87 */
88u_short ip_divert_ignore;
89
90/* Internal variables */
91
92static struct inpcbhead divcb;
93static struct inpcbinfo divcbinfo;
94
95static u_long div_sendspace = DIVSNDQ; /* XXX sysctl ? */
96static u_long div_recvspace = DIVRCVQ; /* XXX sysctl ? */
97
98/* Optimization: have this preinitialized */
99static struct sockaddr_in divsrc = { sizeof(divsrc), AF_INET };
100
101/* Internal functions */
102
103static int div_output(struct socket *so,
104 struct mbuf *m, struct sockaddr *addr, struct mbuf *control);
105
106/*
107 * Initialize divert connection block queue.
108 */
109void
110div_init(void)
111{
112 LIST_INIT(&divcb);
113 divcbinfo.listhead = &divcb;
114 /*
115 * XXX We don't use the hash list for divert IP, but it's easier
116 * to allocate a one entry hash list than it is to check all
117 * over the place for hashbase == NULL.
118 */
119 divcbinfo.hashbase = hashinit(1, M_PCB, &divcbinfo.hashmask);
120 divcbinfo.porthashbase = hashinit(1, M_PCB, &divcbinfo.porthashmask);
121 divcbinfo.ipi_zone = zinit("divcb", sizeof(struct inpcb),
122 maxsockets, ZONE_INTERRUPT, 0);
123}
124
125/*
126 * Setup generic address and protocol structures
127 * for div_input routine, then pass them along with
128 * mbuf chain. ip->ip_len is assumed to have had
129 * the header length (hlen) subtracted out already.
130 * We tell whether the packet was incoming or outgoing
131 * by seeing if hlen == 0, which is a hack.
132 */
133void
134div_input(struct mbuf *m, int hlen)
135{
136 struct ip *ip;
137 struct inpcb *inp;
138 struct socket *sa;
139
140 /* Sanity check */
141 if (ip_divert_port == 0)
142 panic("div_input: port is 0");
143
144 /* Assure header */
145 if (m->m_len < sizeof(struct ip) &&
146 (m = m_pullup(m, sizeof(struct ip))) == 0) {
147 return;
148 }
149 ip = mtod(m, struct ip *);
150
151 /* Record divert port */
152 divsrc.sin_port = htons(ip_divert_port);
153
154 /* Restore packet header fields */
155 ip->ip_len += hlen;
156 HTONS(ip->ip_len);
157 HTONS(ip->ip_off);
158
159 /* Record receive interface address, if any */
160 divsrc.sin_addr.s_addr = 0;
161 if (hlen) {
162 struct ifaddr *ifa;
163
164#ifdef DIAGNOSTIC
165 /* Sanity check */
166 if (!(m->m_flags & M_PKTHDR))
167 panic("div_input: no pkt hdr");
168#endif
169
170 /* More fields affected by ip_input() */
171 HTONS(ip->ip_id);
172
173 /* Find IP address for recieve interface */
174 for (ifa = m->m_pkthdr.rcvif->if_addrhead.tqh_first;
175 ifa != NULL; ifa = ifa->ifa_link.tqe_next) {
176 if (ifa->ifa_addr == NULL)
177 continue;
178 if (ifa->ifa_addr->sa_family != AF_INET)
179 continue;
180 divsrc.sin_addr =
181 ((struct sockaddr_in *) ifa->ifa_addr)->sin_addr;
182 break;
183 }
184 }
185
186 /* Put packet on socket queue, if any */
187 sa = NULL;
188 for (inp = divcb.lh_first; inp != NULL; inp = inp->inp_list.le_next) {
189 if (inp->inp_lport == htons(ip_divert_port))
190 sa = inp->inp_socket;
191 }
192 if (sa) {
193 if (sbappendaddr(&sa->so_rcv, (struct sockaddr *)&divsrc,
194 m, (struct mbuf *)0) == 0)
195 m_freem(m);
196 else
197 sorwakeup(sa);
198 } else {
199 m_freem(m);
200 ipstat.ips_noproto++;
201 ipstat.ips_delivered--;
202 }
203}
204
205/*
206 * Deliver packet back into the IP processing machinery.
207 *
208 * If no address specified, or address is 0.0.0.0, send to ip_output();
209 * otherwise, send to ip_input() and mark as having been received on
210 * the interface with that address.
211 *
212 * If no address specified, or dest port is 0, allow packet to divert
213 * back to this socket; otherwise, don't.
214 */
215static int
216div_output(so, m, addr, control)
217 struct socket *so;
218 register struct mbuf *m;
219 struct sockaddr *addr;
220 struct mbuf *control;
221{
222 register struct inpcb *const inp = sotoinpcb(so);
223 register struct ip *const ip = mtod(m, struct ip *);
224 struct sockaddr_in *sin = NULL;
225 int error = 0;
226
227 if (control)
228 m_freem(control); /* XXX */
229 if (addr)
230 sin = (struct sockaddr_in *)addr;
231
232 /* Loopback avoidance option */
233 ip_divert_ignore = ntohs(inp->inp_lport);
234
235 /* Reinject packet into the system as incoming or outgoing */
236 if (!sin || sin->sin_addr.s_addr == 0) {
237 /* Don't allow both user specified and setsockopt options,
238 and don't allow packet length sizes that will crash */
239 if (((ip->ip_hl != (sizeof (*ip) >> 2)) && inp->inp_options) ||
240 ((u_short)ntohs(ip->ip_len) > m->m_pkthdr.len)) {
241 error = EINVAL;
242 goto cantsend;
243 }
244
245 /* Convert fields to host order for ip_output() */
246 NTOHS(ip->ip_len);
247 NTOHS(ip->ip_off);
248
249 /* Send packet to output processing */
250 ipstat.ips_rawout++; /* XXX */
251 error = ip_output(m, inp->inp_options, &inp->inp_route,
252 (so->so_options & SO_DONTROUTE) |
253 IP_ALLOWBROADCAST | IP_RAWOUTPUT, inp->inp_moptions);
254 } else {
255 struct ifaddr *ifa;
256
257 /* Find receive interface with the given IP address */
258 sin->sin_port = 0;
259 if ((ifa = ifa_ifwithaddr((struct sockaddr *) sin)) == 0) {
260 error = EADDRNOTAVAIL;
261 goto cantsend;
262 }
263 m->m_pkthdr.rcvif = ifa->ifa_ifp;
264
265 /* Send packet to input processing */
266 ip_input(m);
267 }
268
269 /* Reset for next time (and other packets) */
270 ip_divert_ignore = 0;
271 return error;
272
273cantsend:
274 ip_divert_ignore = 0;
275 m_freem(m);
276 return error;
277}
278
279static int
280div_attach(struct socket *so, int proto, struct proc *p)
281{
282 struct inpcb *inp;
283 int error, s;
284
285 inp = sotoinpcb(so);
286 if (inp)
287 panic("div_attach");
288 if (p && (error = suser(p->p_ucred, &p->p_acflag)) != 0)
289 return error;
290
291 s = splnet();
292 error = in_pcballoc(so, &divcbinfo, p);
293 splx(s);
294 if (error)
295 return error;
296 error = soreserve(so, div_sendspace, div_recvspace);
297 if (error)
298 return error;
299 inp = (struct inpcb *)so->so_pcb;
300 inp->inp_ip_p = proto;
301 inp->inp_flags |= INP_HDRINCL;
302 /* The socket is always "connected" because
303 we always know "where" to send the packet */
304 so->so_state |= SS_ISCONNECTED;
305 return 0;
306}
307
308static int
309div_detach(struct socket *so)
310{
311 struct inpcb *inp;
312
313 inp = sotoinpcb(so);
314 if (inp == 0)
315 panic("div_detach");
316 in_pcbdetach(inp);
317 return 0;
318}
319
320static int
321div_abort(struct socket *so)
322{
323 soisdisconnected(so);
324 return div_detach(so);
325}
326
327static int
328div_disconnect(struct socket *so)
329{
330 if ((so->so_state & SS_ISCONNECTED) == 0)
331 return ENOTCONN;
332 return div_abort(so);
333}
334
335static int
336div_bind(struct socket *so, struct sockaddr *nam, struct proc *p)
337{
338 struct inpcb *inp;
339 int s;
340 int error;
341
342 s = splnet();
343 inp = sotoinpcb(so);
344 error = in_pcbbind(inp, nam, p);
345 splx(s);
346 return 0;
347}
348
349static int
350div_shutdown(struct socket *so)
351{
352 socantsendmore(so);
353 return 0;
354}
355
356static int
357div_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam,
358 struct mbuf *control, struct proc *p)
359{
360 /* Packet must have a header (but that's about it) */
361 if (m->m_len < sizeof (struct ip) ||
362 (m = m_pullup(m, sizeof (struct ip))) == 0) {
363 ipstat.ips_toosmall++;
364 m_freem(m);
365 return EINVAL;
366 }
367
368 /* Send packet */
369 return div_output(so, m, nam, control);
370}
371
372struct pr_usrreqs div_usrreqs = {
373 div_abort, pru_accept_notsupp, div_attach, div_bind,
374 pru_connect_notsupp, pru_connect2_notsupp, in_control, div_detach,
375 div_disconnect, pru_listen_notsupp, in_setpeeraddr, pru_rcvd_notsupp,
376 pru_rcvoob_notsupp, div_send, pru_sense_null, div_shutdown,
377 in_setsockaddr, sosend, soreceive, sopoll
378};