tcpdump/print-ip6.c
Francois-Xavier Le Bail f6ae03e299 IPv6: Print some header fields, even if the header is incomplete
Add a test file with a 39-byte header.

Moreover:
Fix an indentation.
2023-08-27 13:28:12 +02:00

508 lines
13 KiB
C

/*
* Copyright (c) 1988, 1989, 1990, 1991, 1992, 1993, 1994
* The Regents of the University of California. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that: (1) source code distributions
* retain the above copyright notice and this paragraph in its entirety, (2)
* distributions including binary code include the above copyright notice and
* this paragraph in its entirety in the documentation or other materials
* provided with the distribution, and (3) all advertising materials mentioning
* features or use of this software display the following acknowledgement:
* ``This product includes software developed by the University of California,
* Lawrence Berkeley Laboratory and its contributors.'' Neither the name of
* the University nor the names of its contributors may be used to endorse
* or promote products derived from this software without specific prior
* written permission.
* THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
* WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
*/
/* \summary: IPv6 printer */
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
#include "netdissect-stdinc.h"
#include <string.h>
#include "netdissect.h"
#include "addrtoname.h"
#include "extract.h"
#include "ip6.h"
#include "ipproto.h"
/*
* If routing headers are presend and valid, set dst to the final destination.
* Otherwise, set it to the IPv6 destination.
*
* This is used for UDP and TCP pseudo-header in the checksum
* calculation.
*/
static void
ip6_finddst(netdissect_options *ndo, nd_ipv6 *dst,
const struct ip6_hdr *ip6)
{
const u_char *cp;
u_int advance;
u_int nh;
const void *dst_addr;
const struct ip6_rthdr *dp;
const struct ip6_rthdr0 *dp0;
const struct ip6_srh *srh;
const u_char *p;
int i, len;
cp = (const u_char *)ip6;
advance = sizeof(struct ip6_hdr);
nh = GET_U_1(ip6->ip6_nxt);
dst_addr = (const void *)ip6->ip6_dst;
while (cp < ndo->ndo_snapend) {
cp += advance;
switch (nh) {
case IPPROTO_HOPOPTS:
case IPPROTO_DSTOPTS:
case IPPROTO_MOBILITY_OLD:
case IPPROTO_MOBILITY:
/*
* These have a header length byte, following
* the next header byte, giving the length of
* the header, in units of 8 octets, excluding
* the first 8 octets.
*/
advance = (GET_U_1(cp + 1) + 1) << 3;
nh = GET_U_1(cp);
break;
case IPPROTO_FRAGMENT:
/*
* The byte following the next header byte is
* marked as reserved, and the header is always
* the same size.
*/
advance = sizeof(struct ip6_frag);
nh = GET_U_1(cp);
break;
case IPPROTO_ROUTING:
/*
* OK, we found it.
*/
dp = (const struct ip6_rthdr *)cp;
ND_TCHECK_SIZE(dp);
len = GET_U_1(dp->ip6r_len);
switch (GET_U_1(dp->ip6r_type)) {
case IPV6_RTHDR_TYPE_0:
case IPV6_RTHDR_TYPE_2: /* Mobile IPv6 ID-20 */
dp0 = (const struct ip6_rthdr0 *)dp;
if (len % 2 == 1)
goto trunc;
len >>= 1;
p = (const u_char *) dp0->ip6r0_addr;
for (i = 0; i < len; i++) {
ND_TCHECK_16(p);
dst_addr = (const void *)p;
p += 16;
}
break;
case IPV6_RTHDR_TYPE_4:
/* IPv6 Segment Routing Header (SRH) */
srh = (const struct ip6_srh *)dp;
if (len % 2 == 1)
goto trunc;
p = (const u_char *) srh->srh_segments;
/*
* The list of segments are encoded in the reverse order.
* Accordingly, the final DA is encoded in srh_segments[0]
*/
ND_TCHECK_16(p);
dst_addr = (const void *)p;
break;
default:
break;
}
/*
* Only one routing header to a customer.
*/
goto done;
case IPPROTO_AH:
case IPPROTO_ESP:
case IPPROTO_IPCOMP:
default:
/*
* AH and ESP are, in the RFCs that describe them,
* described as being "viewed as an end-to-end
* payload" "in the IPv6 context, so that they
* "should appear after hop-by-hop, routing, and
* fragmentation extension headers". We assume
* that's the case, and stop as soon as we see
* one. (We can't handle an ESP header in
* the general case anyway, as its length depends
* on the encryption algorithm.)
*
* IPComp is also "viewed as an end-to-end
* payload" "in the IPv6 context".
*
* All other protocols are assumed to be the final
* protocol.
*/
goto done;
}
}
done:
trunc:
GET_CPY_BYTES(dst, dst_addr, sizeof(nd_ipv6));
}
/*
* Compute a V6-style checksum by building a pseudoheader.
*/
uint16_t
nextproto6_cksum(netdissect_options *ndo,
const struct ip6_hdr *ip6, const uint8_t *data,
u_int len, u_int covlen, uint8_t next_proto)
{
struct {
nd_ipv6 ph_src;
nd_ipv6 ph_dst;
uint32_t ph_len;
uint8_t ph_zero[3];
uint8_t ph_nxt;
} ph;
struct cksum_vec vec[2];
u_int nh;
/* pseudo-header */
memset(&ph, 0, sizeof(ph));
GET_CPY_BYTES(&ph.ph_src, ip6->ip6_src, sizeof(nd_ipv6));
nh = GET_U_1(ip6->ip6_nxt);
switch (nh) {
case IPPROTO_HOPOPTS:
case IPPROTO_DSTOPTS:
case IPPROTO_MOBILITY_OLD:
case IPPROTO_MOBILITY:
case IPPROTO_FRAGMENT:
case IPPROTO_ROUTING:
/*
* The next header is either a routing header or a header
* after which there might be a routing header, so scan
* for a routing header.
*/
ip6_finddst(ndo, &ph.ph_dst, ip6);
break;
default:
GET_CPY_BYTES(&ph.ph_dst, ip6->ip6_dst, sizeof(nd_ipv6));
break;
}
ph.ph_len = htonl(len);
ph.ph_nxt = next_proto;
vec[0].ptr = (const uint8_t *)(void *)&ph;
vec[0].len = sizeof(ph);
vec[1].ptr = data;
vec[1].len = covlen;
return in_cksum(vec, 2);
}
/*
* print an IP6 datagram.
*/
void
ip6_print(netdissect_options *ndo, const u_char *bp, u_int length)
{
const struct ip6_hdr *ip6;
int advance;
u_int len;
u_int total_advance;
const u_char *cp;
uint32_t payload_len;
uint8_t ph, nh;
int fragmented = 0;
u_int flow;
int found_extension_header;
int found_jumbo;
int found_hbh;
ndo->ndo_protocol = "ip6";
ip6 = (const struct ip6_hdr *)bp;
if (!ndo->ndo_eflag) {
nd_print_protocol_caps(ndo);
ND_PRINT(" ");
}
ND_ICHECK_ZU(length, <, sizeof (struct ip6_hdr));
ND_ICHECKMSG_U("version", IP6_VERSION(ip6), !=, 6);
payload_len = GET_BE_U_2(ip6->ip6_plen);
/*
* RFC 1883 says:
*
* The Payload Length field in the IPv6 header must be set to zero
* in every packet that carries the Jumbo Payload option. If a
* packet is received with a valid Jumbo Payload option present and
* a non-zero IPv6 Payload Length field, an ICMP Parameter Problem
* message, Code 0, should be sent to the packet's source, pointing
* to the Option Type field of the Jumbo Payload option.
*
* Later versions of the IPv6 spec don't discuss the Jumbo Payload
* option.
*
* If the payload length is 0, we temporarily just set the total
* length to the remaining data in the packet (which, for Ethernet,
* could include frame padding, but if it's a Jumbo Payload frame,
* it shouldn't even be sendable over Ethernet, so we don't worry
* about that), so we can process the extension headers in order
* to *find* a Jumbo Payload hop-by-hop option and, when we've
* processed all the extension headers, check whether we found
* a Jumbo Payload option, and fail if we haven't.
*/
if (payload_len != 0) {
len = payload_len + sizeof(struct ip6_hdr);
if (len > length) {
ND_PRINT("[payload+header length %u > length %u]",
len, length);
nd_print_invalid(ndo);
ND_PRINT(" ");
}
} else
len = length + sizeof(struct ip6_hdr);
ph = 255;
nh = GET_U_1(ip6->ip6_nxt);
if (ndo->ndo_vflag) {
flow = GET_BE_U_4(ip6->ip6_flow);
ND_PRINT("(");
/* RFC 2460 */
if (flow & 0x0ff00000)
ND_PRINT("class 0x%02x, ", (flow & 0x0ff00000) >> 20);
if (flow & 0x000fffff)
ND_PRINT("flowlabel 0x%05x, ", flow & 0x000fffff);
ND_PRINT("hlim %u, next-header %s (%u) payload length: %u) ",
GET_U_1(ip6->ip6_hlim),
tok2str(ipproto_values,"unknown",nh),
nh,
payload_len);
}
ND_TCHECK_SIZE(ip6);
/*
* Cut off the snapshot length to the end of the IP payload.
*/
if (!nd_push_snaplen(ndo, bp, len)) {
(*ndo->ndo_error)(ndo, S_ERR_ND_MEM_ALLOC,
"%s: can't push snaplen on buffer stack", __func__);
}
cp = (const u_char *)ip6;
advance = sizeof(struct ip6_hdr);
total_advance = 0;
/* Process extension headers */
found_extension_header = 0;
found_jumbo = 0;
found_hbh = 0;
while (cp < ndo->ndo_snapend && advance > 0) {
if (len < (u_int)advance)
goto trunc;
cp += advance;
len -= advance;
total_advance += advance;
if (cp == (const u_char *)(ip6 + 1) &&
nh != IPPROTO_TCP && nh != IPPROTO_UDP &&
nh != IPPROTO_DCCP && nh != IPPROTO_SCTP) {
ND_PRINT("%s > %s: ", GET_IP6ADDR_STRING(ip6->ip6_src),
GET_IP6ADDR_STRING(ip6->ip6_dst));
}
switch (nh) {
case IPPROTO_HOPOPTS:
/*
* The Hop-by-Hop Options header, when present,
* must immediately follow the IPv6 header (RFC 8200)
*/
if (found_hbh == 1) {
ND_PRINT("[The Hop-by-Hop Options header was already found]");
nd_print_invalid(ndo);
return;
}
if (ph != 255) {
ND_PRINT("[The Hop-by-Hop Options header don't follow the IPv6 header]");
nd_print_invalid(ndo);
return;
}
advance = hbhopt_process(ndo, cp, &found_jumbo, &payload_len);
if (payload_len == 0 && found_jumbo == 0) {
ND_PRINT("[No valid Jumbo Payload Hop-by-Hop option found]");
nd_print_invalid(ndo);
return;
}
if (advance < 0) {
nd_pop_packet_info(ndo);
return;
}
found_extension_header = 1;
found_hbh = 1;
nh = GET_U_1(cp);
break;
case IPPROTO_DSTOPTS:
advance = dstopt_process(ndo, cp);
if (advance < 0) {
nd_pop_packet_info(ndo);
return;
}
found_extension_header = 1;
nh = GET_U_1(cp);
break;
case IPPROTO_FRAGMENT:
advance = frag6_print(ndo, cp, (const u_char *)ip6);
if (advance < 0 || ndo->ndo_snapend <= cp + advance) {
nd_pop_packet_info(ndo);
return;
}
found_extension_header = 1;
nh = GET_U_1(cp);
fragmented = 1;
break;
case IPPROTO_MOBILITY_OLD:
case IPPROTO_MOBILITY:
/*
* XXX - we don't use "advance"; RFC 3775 says that
* the next header field in a mobility header
* should be IPPROTO_NONE, but speaks of
* the possibility of a future extension in
* which payload can be piggybacked atop a
* mobility header.
*/
advance = mobility_print(ndo, cp, (const u_char *)ip6);
if (advance < 0) {
nd_pop_packet_info(ndo);
return;
}
found_extension_header = 1;
nh = GET_U_1(cp);
nd_pop_packet_info(ndo);
return;
case IPPROTO_ROUTING:
ND_TCHECK_1(cp);
advance = rt6_print(ndo, cp, (const u_char *)ip6);
if (advance < 0) {
nd_pop_packet_info(ndo);
return;
}
found_extension_header = 1;
nh = GET_U_1(cp);
break;
default:
/*
* Not an extension header; hand off to the
* IP protocol demuxer.
*/
if (found_jumbo) {
/*
* We saw a Jumbo Payload option.
* Set the length to the payload length
* plus the IPv6 header length, and
* change the snapshot length accordingly.
*
* But make sure it's not shorter than
* the total number of bytes we've
* processed so far.
*/
len = payload_len + sizeof(struct ip6_hdr);
if (len < total_advance)
goto trunc;
if (len > length) {
ND_PRINT("[payload+header length %u > length %u]",
len, length);
nd_print_invalid(ndo);
ND_PRINT(" ");
}
nd_change_snaplen(ndo, bp, len);
/*
* Now subtract the length of the IPv6
* header plus extension headers to get
* the payload length.
*/
len -= total_advance;
} else {
/*
* We didn't see a Jumbo Payload option;
* was the payload length zero?
*/
if (payload_len == 0) {
/*
* Yes. If we found an extension
* header, treat that as a truncated
* packet header, as there was
* no payload to contain an
* extension header.
*/
if (found_extension_header)
goto trunc;
/*
* OK, we didn't see any extension
* header, but that means we have
* no payload, so set the length
* to the IPv6 header length,
* and change the snapshot length
* accordingly.
*/
len = sizeof(struct ip6_hdr);
nd_change_snaplen(ndo, bp, len);
/*
* Now subtract the length of
* the IPv6 header plus extension
* headers (there weren't any, so
* that's just the IPv6 header
* length) to get the payload length.
*/
len -= total_advance;
}
}
ip_demux_print(ndo, cp, len, 6, fragmented,
GET_U_1(ip6->ip6_hlim), nh, bp);
nd_pop_packet_info(ndo);
return;
}
ph = nh;
/* ndo_protocol reassignment after xxx_print() calls */
ndo->ndo_protocol = "ip6";
}
nd_pop_packet_info(ndo);
return;
trunc:
nd_print_trunc(ndo);
return;
invalid:
nd_print_invalid(ndo);
}