2 * Copyright (c) 1982, 1986, 1988, 1993
3 * The Regents of the University of California. All rights reserved.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
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. Neither the name of the University nor the names of its contributors
14 * may be used to endorse or promote products derived from this software
15 * without specific prior written permission.
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * @(#)ip_icmp.c 8.2 (Berkeley) 1/4/94
30 * ip_icmp.c,v 1.7 1995/05/30 08:09:42 rgrimes Exp
33 #include "qemu/osdep.h"
37 /* The message sent when emulating PING */
38 /* Be nice and tell them it's just a pseudo-ping packet */
39 static const char icmp_ping_msg
[] = "This is a pseudo-PING packet used by Slirp to emulate ICMP ECHO-REQUEST packets.\n";
41 /* list of actions for icmp_send_error() on RX of an icmp message */
42 static const int icmp_flush
[19] = {
43 /* ECHO REPLY (0) */ 0,
46 /* DEST UNREACH (3) */ 1,
47 /* SOURCE QUENCH (4)*/ 1,
52 /* ROUTERADVERT (9) */ 1,
53 /* ROUTERSOLICIT (10) */ 1,
54 /* TIME EXCEEDED (11) */ 1,
55 /* PARAMETER PROBLEM (12) */ 1,
56 /* TIMESTAMP (13) */ 0,
57 /* TIMESTAMP REPLY (14) */ 0,
59 /* INFO REPLY (16) */ 0,
60 /* ADDR MASK (17) */ 0,
61 /* ADDR MASK REPLY (18) */ 0
64 void icmp_init(Slirp
*slirp
)
66 slirp
->icmp
.so_next
= slirp
->icmp
.so_prev
= &slirp
->icmp
;
67 slirp
->icmp_last_so
= &slirp
->icmp
;
70 void icmp_cleanup(Slirp
*slirp
)
72 while (slirp
->icmp
.so_next
!= &slirp
->icmp
) {
73 icmp_detach(slirp
->icmp
.so_next
);
77 static int icmp_send(struct socket
*so
, struct mbuf
*m
, int hlen
)
79 struct ip
*ip
= mtod(m
, struct ip
*);
80 struct sockaddr_in addr
;
82 so
->s
= qemu_socket(AF_INET
, SOCK_DGRAM
, IPPROTO_ICMP
);
88 so
->so_faddr
= ip
->ip_dst
;
89 so
->so_laddr
= ip
->ip_src
;
90 so
->so_iptos
= ip
->ip_tos
;
91 so
->so_type
= IPPROTO_ICMP
;
92 so
->so_state
= SS_ISFCONNECTED
;
93 so
->so_expire
= curtime
+ SO_EXPIRE
;
95 addr
.sin_family
= AF_INET
;
96 addr
.sin_addr
= so
->so_faddr
;
98 insque(so
, &so
->slirp
->icmp
);
100 if (sendto(so
->s
, m
->m_data
+ hlen
, m
->m_len
- hlen
, 0,
101 (struct sockaddr
*)&addr
, sizeof(addr
)) == -1) {
102 DEBUG_MISC((dfd
, "icmp_input icmp sendto tx errno = %d-%s\n",
103 errno
, strerror(errno
)));
104 icmp_send_error(m
, ICMP_UNREACH
, ICMP_UNREACH_NET
, 0, strerror(errno
));
111 void icmp_detach(struct socket
*so
)
118 * Process a received ICMP message.
121 icmp_input(struct mbuf
*m
, int hlen
)
123 register struct icmp
*icp
;
124 register struct ip
*ip
=mtod(m
, struct ip
*);
125 int icmplen
=ip
->ip_len
;
126 Slirp
*slirp
= m
->slirp
;
128 DEBUG_CALL("icmp_input");
129 DEBUG_ARG("m = %p", m
);
130 DEBUG_ARG("m_len = %d", m
->m_len
);
133 * Locate icmp structure in mbuf, and check
134 * that its not corrupted and of at least minimum length.
136 if (icmplen
< ICMP_MINLEN
) { /* min 8 bytes payload */
144 icp
= mtod(m
, struct icmp
*);
145 if (cksum(m
, icmplen
)) {
151 DEBUG_ARG("icmp_type = %d", icp
->icmp_type
);
152 switch (icp
->icmp_type
) {
154 ip
->ip_len
+= hlen
; /* since ip_input subtracts this */
155 if (ip
->ip_dst
.s_addr
== slirp
->vhost_addr
.s_addr
||
156 ip
->ip_dst
.s_addr
== slirp
->vnameserver_addr
.s_addr
) {
158 } else if (slirp
->restricted
) {
162 struct sockaddr_storage addr
;
163 if ((so
= socreate(slirp
)) == NULL
) goto freeit
;
164 if (icmp_send(so
, m
, hlen
) == 0) {
167 if (udp_attach(so
, AF_INET
) == -1) {
168 DEBUG_MISC((dfd
,"icmp_input udp_attach errno = %d-%s\n",
169 errno
,strerror(errno
)));
175 so
->so_ffamily
= AF_INET
;
176 so
->so_faddr
= ip
->ip_dst
;
177 so
->so_fport
= htons(7);
178 so
->so_lfamily
= AF_INET
;
179 so
->so_laddr
= ip
->ip_src
;
180 so
->so_lport
= htons(9);
181 so
->so_iptos
= ip
->ip_tos
;
182 so
->so_type
= IPPROTO_ICMP
;
183 so
->so_state
= SS_ISFCONNECTED
;
185 /* Send the packet */
187 sotranslate_out(so
, &addr
);
189 if(sendto(so
->s
, icmp_ping_msg
, strlen(icmp_ping_msg
), 0,
190 (struct sockaddr
*)&addr
, sockaddr_size(&addr
)) == -1) {
191 DEBUG_MISC((dfd
,"icmp_input udp sendto tx errno = %d-%s\n",
192 errno
,strerror(errno
)));
193 icmp_send_error(m
, ICMP_UNREACH
, ICMP_UNREACH_NET
, 0, strerror(errno
));
196 } /* if ip->ip_dst.s_addr == alias_addr.s_addr */
199 /* XXX? report error? close socket? */
202 case ICMP_SOURCEQUENCH
:
214 /* m is m_free()'d xor put in a socket xor or given to ip_send */
220 * Send an ICMP message in response to a situation
222 * RFC 1122: 3.2.2 MUST send at least the IP header and 8 bytes of header. MAY send more (we do).
223 * MUST NOT change this header information.
224 * MUST NOT reply to a multicast/broadcast IP address.
225 * MUST NOT reply to a multicast/broadcast MAC address.
226 * MUST reply to only the first fragment.
229 * Send ICMP_UNREACH back to the source regarding msrc.
230 * mbuf *msrc is used as a template, but is NOT m_free()'d.
231 * It is reported as the bad ip packet. The header should
232 * be fully correct and in host byte order.
233 * ICMP fragmentation is illegal. All machines must accept 576 bytes in one
234 * packet. The maximum payload is 576-20(ip hdr)-8(icmp hdr)=548
237 #define ICMP_MAXDATALEN (IP_MSS-28)
239 icmp_send_error(struct mbuf
*msrc
, u_char type
, u_char code
, int minsize
,
242 unsigned hlen
, shlen
, s_ip_len
;
243 register struct ip
*ip
;
244 register struct icmp
*icp
;
245 register struct mbuf
*m
;
247 DEBUG_CALL("icmp_send_error");
248 DEBUG_ARG("msrc = %p", msrc
);
249 DEBUG_ARG("msrc_len = %d", msrc
->m_len
);
251 if(type
!=ICMP_UNREACH
&& type
!=ICMP_TIMXCEED
) goto end_error
;
254 if(!msrc
) goto end_error
;
255 ip
= mtod(msrc
, struct ip
*);
257 { char bufa
[20], bufb
[20];
258 strcpy(bufa
, inet_ntoa(ip
->ip_src
));
259 strcpy(bufb
, inet_ntoa(ip
->ip_dst
));
260 DEBUG_MISC((dfd
, " %.16s to %.16s\n", bufa
, bufb
));
263 if(ip
->ip_off
& IP_OFFMASK
) goto end_error
; /* Only reply to fragment 0 */
265 /* Do not reply to source-only IPs */
266 if ((ip
->ip_src
.s_addr
& htonl(~(0xf << 28))) == 0) {
270 shlen
=ip
->ip_hl
<< 2;
272 if(ip
->ip_p
== IPPROTO_ICMP
) {
273 icp
= (struct icmp
*)((char *)ip
+ shlen
);
275 * Assume any unknown ICMP type is an error. This isn't
276 * specified by the RFC, but think about it..
278 if(icp
->icmp_type
>18 || icmp_flush
[icp
->icmp_type
]) goto end_error
;
282 m
= m_get(msrc
->slirp
);
288 new_m_size
=sizeof(struct ip
)+ICMP_MINLEN
+msrc
->m_len
+ICMP_MAXDATALEN
;
289 if(new_m_size
>m
->m_size
) m_inc(m
, new_m_size
);
291 memcpy(m
->m_data
, msrc
->m_data
, msrc
->m_len
);
292 m
->m_len
= msrc
->m_len
; /* copy msrc to m */
294 /* make the header of the reply packet */
295 ip
= mtod(m
, struct ip
*);
296 hlen
= sizeof(struct ip
); /* no options in reply */
302 icp
= mtod(m
, struct icmp
*);
304 if(minsize
) s_ip_len
=shlen
+ICMP_MINLEN
; /* return header+8b only */
305 else if(s_ip_len
>ICMP_MAXDATALEN
) /* maximum size */
306 s_ip_len
=ICMP_MAXDATALEN
;
308 m
->m_len
=ICMP_MINLEN
+s_ip_len
; /* 8 bytes ICMP header */
310 /* min. size = 8+sizeof(struct ip)+8 */
312 icp
->icmp_type
= type
;
313 icp
->icmp_code
= code
;
317 memcpy(&icp
->icmp_ip
, msrc
->m_data
, s_ip_len
); /* report the ip packet */
318 HTONS(icp
->icmp_ip
.ip_len
);
319 HTONS(icp
->icmp_ip
.ip_id
);
320 HTONS(icp
->icmp_ip
.ip_off
);
323 if(message
) { /* DEBUG : append message to ICMP packet */
326 message_len
=strlen(message
);
327 if(message_len
>ICMP_MAXDATALEN
) message_len
=ICMP_MAXDATALEN
;
328 cpnt
=(char *)m
->m_data
+m
->m_len
;
329 memcpy(cpnt
, message
, message_len
);
330 m
->m_len
+=message_len
;
335 icp
->icmp_cksum
= cksum(m
, m
->m_len
);
341 ip
->ip_hl
= hlen
>> 2;
342 ip
->ip_len
= m
->m_len
;
344 ip
->ip_tos
=((ip
->ip_tos
& 0x1E) | 0xC0); /* high priority for errors */
347 ip
->ip_p
= IPPROTO_ICMP
;
348 ip
->ip_dst
= ip
->ip_src
; /* ip addresses */
349 ip
->ip_src
= m
->slirp
->vhost_addr
;
351 (void ) ip_output((struct socket
*)NULL
, m
);
356 #undef ICMP_MAXDATALEN
359 * Reflect the ip packet back to the source
362 icmp_reflect(struct mbuf
*m
)
364 register struct ip
*ip
= mtod(m
, struct ip
*);
365 int hlen
= ip
->ip_hl
<< 2;
366 int optlen
= hlen
- sizeof(struct ip
);
367 register struct icmp
*icp
;
370 * Send an icmp packet back to the ip level,
371 * after supplying a checksum.
375 icp
= mtod(m
, struct icmp
*);
377 icp
->icmp_type
= ICMP_ECHOREPLY
;
379 icp
->icmp_cksum
= cksum(m
, ip
->ip_len
- hlen
);
387 * Strip out original options by copying rest of first
388 * mbuf's data back, and adjust the IP length.
390 memmove((caddr_t
)(ip
+ 1), (caddr_t
)ip
+ hlen
,
391 (unsigned )(m
->m_len
- hlen
));
393 ip
->ip_hl
= hlen
>> 2;
394 ip
->ip_len
-= optlen
;
400 struct in_addr icmp_dst
;
401 icmp_dst
= ip
->ip_dst
;
402 ip
->ip_dst
= ip
->ip_src
;
403 ip
->ip_src
= icmp_dst
;
406 (void ) ip_output((struct socket
*)NULL
, m
);
409 void icmp_receive(struct socket
*so
)
411 struct mbuf
*m
= so
->so_m
;
412 struct ip
*ip
= mtod(m
, struct ip
*);
413 int hlen
= ip
->ip_hl
<< 2;
420 icp
= mtod(m
, struct icmp
*);
423 len
= qemu_recv(so
->s
, icp
, M_ROOM(m
), 0);
425 * The behavior of reading SOCK_DGRAM+IPPROTO_ICMP sockets is inconsistent
426 * between host OSes. On Linux, only the ICMP header and payload is
427 * included. On macOS/Darwin, the socket acts like a raw socket and
428 * includes the IP header as well. On other BSDs, SOCK_DGRAM+IPPROTO_ICMP
429 * sockets aren't supported at all, so we treat them like raw sockets. It
430 * isn't possible to detect this difference at runtime, so we must use an
431 * #ifdef to determine if we need to remove the IP header.
434 if (len
>= sizeof(struct ip
)) {
435 struct ip
*inner_ip
= mtod(m
, struct ip
*);
436 int inner_hlen
= inner_ip
->ip_hl
<< 2;
437 if (inner_hlen
> len
) {
442 memmove(icp
, (unsigned char *)icp
+ inner_hlen
, len
);
454 if (len
== -1 || len
== 0) {
455 if (errno
== ENETUNREACH
) {
456 error_code
= ICMP_UNREACH_NET
;
458 error_code
= ICMP_UNREACH_HOST
;
460 DEBUG_MISC((dfd
, " udp icmp rx errno = %d-%s\n", errno
,
462 icmp_send_error(so
->so_m
, ICMP_UNREACH
, error_code
, 0, strerror(errno
));
464 icmp_reflect(so
->so_m
);
465 so
->so_m
= NULL
; /* Don't m_free() it again! */