]> The Tcpdump Group git mirrors - tcpdump/log
tcpdump
21 years agoPrint the subtype of management frames *before* parsing the body, so
guy [Thu, 27 Nov 2003 02:31:15 +0000 (02:31 +0000)] 
Print the subtype of management frames *before* parsing the body, so
that if there's an error parsing the body, we at least see the type (and
so that the code to print the subtype isn't duplicated in every
body-printing routine).

21 years ago135 appears to be the current IP protocol number for mobility headers;
guy [Mon, 24 Nov 2003 20:31:22 +0000 (20:31 +0000)] 
135 appears to be the current IP protocol number for mobility headers;
support it, as well as 62 (which presumably was one that was used
before, even though it was already assigned to some protocol named
CFTP).

21 years agoPropagate Gerald Combs' man page fix to document "<<" and ">>".
guy [Sun, 23 Nov 2003 23:43:41 +0000 (23:43 +0000)] 
Propagate Gerald Combs' man page fix to document "<<" and ">>".

21 years agoDon't check for the OSPF opaque traffic engineering TLV type before
guy [Wed, 19 Nov 2003 09:43:47 +0000 (09:43 +0000)] 
Don't check for the OSPF opaque traffic engineering TLV type before
processing the TLV - there might not be any TLVs.  Instead, check before
fetching the type.

21 years agoAdd missing bounds checking to the DVMRP and PIM print routines.
guy [Wed, 19 Nov 2003 09:41:28 +0000 (09:41 +0000)] 
Add missing bounds checking to the DVMRP and PIM print routines.

Instead of checking that there are 8 bytes available at the beginning of
an IGMP packet - which there might not be, e.g. with some DVMRP packets
- do the checks as necessary before fetching values.

21 years agoUse "len", not "payload_len", as we step through the packet; use
guy [Wed, 19 Nov 2003 06:31:10 +0000 (06:31 +0000)] 
Use "len", not "payload_len", as we step through the packet; use
"payload_len" to report the raw value from the header.

21 years agoKeep the remaining payload length in a separate variable from the total
guy [Wed, 19 Nov 2003 06:17:36 +0000 (06:17 +0000)] 
Keep the remaining payload length in a separate variable from the total
payload length, so we can report the latter at the end of dissection.

21 years agoFix a keyboardo.
guy [Wed, 19 Nov 2003 05:36:40 +0000 (05:36 +0000)] 
Fix a keyboardo.

21 years agoFix a typo (as per NetBSD).
guy [Wed, 19 Nov 2003 02:02:00 +0000 (02:02 +0000)] 
Fix a typo (as per NetBSD).

21 years agoFix capitalization of ULTRIX and AppleTalk, as per NetBSD.
guy [Wed, 19 Nov 2003 01:58:58 +0000 (01:58 +0000)] 
Fix capitalization of ULTRIX and AppleTalk, as per NetBSD.

Update the date stamp.

21 years ago"icmp6_rrenum_print()" doesn't use its "icmp6len" argument (and, as that
guy [Wed, 19 Nov 2003 01:28:18 +0000 (01:28 +0000)] 
"icmp6_rrenum_print()" doesn't use its "icmp6len" argument (and, as that
just reflects the payload length, the snapshot length has been trimmed
to correspond to it, so it doesn't need to check it merely to avoid
running past the end of the IPv6 payload); get rid of it.

Pass "icmp6_print()" the IPv6 payload length (i.e., the length of the
ICMPv6 packet), rather than having it compre the length itself.

Make the length argument to "icmp6_cksum()" unsigned, to match the value
passed to it.

Fix the lengths passed to "sctp_print()", "tcp_print()", and
"udp_print()" (we update "len" in the header-processing loop to
correspond to the remaining payload, so we just pass it to those
routines).

21 years agoPut in a long comment explaining what's confusing about this routine,
guy [Wed, 19 Nov 2003 01:09:12 +0000 (01:09 +0000)] 
Put in a long comment explaining what's confusing about this routine,
and explaining that we leave the "unused argument" warning in place as a
reminder that we need to fix this routine someday, when we actually have
a capture against which to test it.

21 years agoHave various routines for printing non-final headers for IPv4/IPv6
guy [Wed, 19 Nov 2003 00:35:43 +0000 (00:35 +0000)] 
Have various routines for printing non-final headers for IPv4/IPv6
return -1 if they run out of data.

Have the IPv4 and IPv6 dissectors check for non-positive return values
from those routines and quit if they see one.

21 years agoFix a signed-vs-unsigned comparison warning.
guy [Wed, 19 Nov 2003 00:19:25 +0000 (00:19 +0000)] 
Fix a signed-vs-unsigned comparison warning.

21 years agoCut off "snapend" at the length of the IPv4 or IPv6 payload, so we don't
guy [Wed, 19 Nov 2003 00:17:01 +0000 (00:17 +0000)] 
Cut off "snapend" at the length of the IPv4 or IPv6 payload, so we don't
run past the end of that payload.

Check that the IPv4 total length isn't less than the header length.

Use "%u", not "%d", to print unsigned values.

Properly update "len" in the header-processing loop for IPv6.

Doing so means we can trust the length passed to the TCP and UDP
dissectors when constructing the IPv6 pseudo-header; do so (but fix the
length we pass to the UDP-over-IPv6 checksum routine).

That length is unsigned; make the corresponding arguments to the TCP and
UDP checksum routines unsigned.

21 years agoGet rid of unused arguments to "dhcp6_print()".
guy [Tue, 18 Nov 2003 23:26:14 +0000 (23:26 +0000)] 
Get rid of unused arguments to "dhcp6_print()".

21 years agoPrint both the name and the description of a link-layer type (the name
guy [Tue, 18 Nov 2003 23:17:27 +0000 (23:17 +0000)] 
Print both the name and the description of a link-layer type (the name
isn't always very suggestive - for example, somebody might think
"EN10MB" is always 10 MB/s, and might not know that "IEEE802" is Token
Ring), using "pcap_datalink_val_to_description()".  Supply our own
"pcap_datalink_val_to_description()" if libpcap doesn't have it (even if
it has "pcap_datalink_name_to_val()").

21 years agoPrint both the name and the description of a link-layer type (the name
guy [Tue, 18 Nov 2003 23:12:10 +0000 (23:12 +0000)] 
Print both the name and the description of a link-layer type (the name
isn't always very suggestive - for example, somebody might think
"EN10MB" is always 10 MB/s, and might not know that "IEEE802" is Token
Ring), using "pcap_datalink_val_to_description()".  Supply our own
"pcap_datalink_val_to_description()" if libpcap doesn't have it (even if
it has "pcap_datalink_name_to_val()").

21 years agoIn MSVC++'s C library, _IOLBF is the same as _IOFBF; use _IONBF instead.
guy [Tue, 18 Nov 2003 08:53:40 +0000 (08:53 +0000)] 
In MSVC++'s C library, _IOLBF is the same as _IOFBF; use _IONBF instead.

Fix the timeout in the "pcap_open_live()" call - I accidentally checked
in a larger value I was using while testing pcap_breakloop().

21 years agoLatest config.guess and config.sub from
guy [Tue, 18 Nov 2003 06:27:17 +0000 (06:27 +0000)] 
Latest config.guess and config.sub from

http://savannah.gnu.org/projects/config/

21 years agoUpdate the RCS tag information for 3.8.
guy [Sun, 16 Nov 2003 10:17:30 +0000 (10:17 +0000)] 
Update the RCS tag information for 3.8.

21 years agoFrom Brian Ginsbach: add the Cray NV1 CPU to the list of CPUs that don't
guy [Sun, 16 Nov 2003 10:14:19 +0000 (10:14 +0000)] 
From Brian Ginsbach: add the Cray NV1 CPU to the list of CPUs that don't
support unaligned accesses.

21 years agoGive it a proper RCS ID.
guy [Sun, 16 Nov 2003 09:57:50 +0000 (09:57 +0000)] 
Give it a proper RCS ID.

21 years agoFrom Brian Ginsbach: add the Cray NV1 CPU to the list of CPUs that don't
guy [Sun, 16 Nov 2003 09:44:37 +0000 (09:44 +0000)] 
From Brian Ginsbach: add the Cray NV1 CPU to the list of CPUs that don't
support unaligned accesses.

21 years agoGive it a proper RCS ID.
guy [Sun, 16 Nov 2003 09:41:56 +0000 (09:41 +0000)] 
Give it a proper RCS ID.

21 years agoInclude "config.h" before including "tcpdump-stdinc.h" in
guy [Sun, 16 Nov 2003 09:29:48 +0000 (09:29 +0000)] 
Include "config.h" before including "tcpdump-stdinc.h" in
"missing/datalinks.c".

21 years agoHave the configure script arrange that the Makefile define _U_
guy [Sun, 16 Nov 2003 08:59:07 +0000 (08:59 +0000)] 
Have the configure script arrange that the Makefile define _U_
appropriately, and that GNUmakefile and the MSVC++ project file define
it apppriately, as we do with libpcap, rather than defining it in
"interface.h".

Undo the rcsid-shuffling and addition of extra #includes, as we no
longer need to arrange that "interface.h" be included before using _U_
in an RCS ID or copyright.

21 years agoHave the configure script arrange that the Makefile define _U_
guy [Sun, 16 Nov 2003 08:51:01 +0000 (08:51 +0000)] 
Have the configure script arrange that the Makefile define _U_
appropriately, and that GNUmakefile and the MSVC++ project file define
it apppriately, as we do with libpcap, rather than defining it in
"interface.h".

Undo the rcsid-shuffling and addition of extra #includes, as we no
longer need to arrange that "interface.h" be included before using _U_
in an RCS ID or copyright.

21 years agoEscape question marks, to avoid warnings about trigraphs.
guy [Sat, 15 Nov 2003 22:34:57 +0000 (22:34 +0000)] 
Escape question marks, to avoid warnings about trigraphs.

21 years agoUse the right format to print a value as 2 hex digits.
guy [Sat, 15 Nov 2003 22:33:28 +0000 (22:33 +0000)] 
Use the right format to print a value as 2 hex digits.

21 years agoFrom Neil Spring:
guy [Sat, 15 Nov 2003 22:28:35 +0000 (22:28 +0000)] 
From Neil Spring:

use "_U_" in the definitions of "rcsid[]", to eliminate
complaints about those variables being unused;

move the definitions after the include of "interface.h", or add
an include of "interface.h", so that "_U_" is defined.

Include "config.h" before including "tcpdump-stdinc.h" in
"missing/datalinks.c".

21 years agoEscape question marks, to avoid warnings about trigraphs.
guy [Sat, 15 Nov 2003 01:08:37 +0000 (01:08 +0000)] 
Escape question marks, to avoid warnings about trigraphs.

21 years agoUse the right format to print a value as 2 hex digits.
guy [Sat, 15 Nov 2003 01:02:02 +0000 (01:02 +0000)] 
Use the right format to print a value as 2 hex digits.

21 years agoFrom Neil Spring:
guy [Sat, 15 Nov 2003 00:39:12 +0000 (00:39 +0000)] 
From Neil Spring:

use "_U_" in the definitions of "rcsid[]", to eliminate
complaints about those variables being unused;

move the definitions after the include of "interface.h", or add
an include of "interface.h", so that "_U_" is defined.

Include "config.h" before including "tcpdump-stdinc.h" in
"missing/datalinks.c".

21 years ago draft CHANGES file for 3.8. tcpdump-3.8-bp
mcr [Thu, 13 Nov 2003 04:06:02 +0000 (04:06 +0000)] 
draft CHANGES file for 3.8.

21 years ago removed people who shouldn't be bugged since they aren't
mcr [Thu, 13 Nov 2003 03:14:55 +0000 (03:14 +0000)] 
removed people who shouldn't be bugged since they aren't
actively committing code now.

21 years ago ignore pcap files which are generated from .puu files.
mcr [Wed, 12 Nov 2003 15:56:22 +0000 (15:56 +0000)] 
ignore pcap files which are generated from .puu files.

21 years agoFix a typo.
guy [Sat, 8 Nov 2003 09:00:51 +0000 (09:00 +0000)] 
Fix a typo.

21 years agodo not attempt to verify the TLV #12 checksum if it is zero - change the way how...
hannes [Wed, 5 Nov 2003 23:11:03 +0000 (23:11 +0000)] 
do not attempt to verify the TLV #12 checksum if it is zero - change the way how the authentication start offset for LSPs is determined

21 years agoadded print-bfd.c to the Windows VC project
risso [Wed, 5 Nov 2003 20:36:16 +0000 (20:36 +0000)] 
added print-bfd.c to the Windows VC project

21 years agostreamline variables and defines# to a common scheme
hannes [Wed, 5 Nov 2003 16:40:57 +0000 (16:40 +0000)] 
streamline variables and defines# to a common scheme

21 years agoIn mDNS, report IN-class records with the "cache flush" bit set as such,
guy [Wed, 5 Nov 2003 06:02:58 +0000 (06:02 +0000)] 
In mDNS, report IN-class records with the "cache flush" bit set as such,
rather than as "Class 32769".

21 years agoCheck whether we have "pcap_breakloop()", and use it if we do - it's a
guy [Tue, 4 Nov 2003 07:29:15 +0000 (07:29 +0000)] 
Check whether we have "pcap_breakloop()", and use it if we do - it's a
safer way to terminate "pcap_loop()" in a signal handler (it just sets a
flag, it doesn't muck with data structures that might have been in the
middle of being updated).

Have "setsignal()" not request SA_RESTART, so that if we call
"pcap_breakloop()" in a signal handler and then return, we don't restart
a call that was waiting for captured packets, we just make that call
EINTR out.

21 years agoPut in the entry for Scott Rose:
guy [Tue, 4 Nov 2003 01:43:20 +0000 (01:43 +0000)] 
Put in the entry for Scott Rose:

The test for DNS message dnssec bit flags in the query and
response are backwards.  The CD bit has meaning in a query, and
the AD bit has meaning in a response.

21 years agodo not attempt to checksum if the LSP has been purged
hannes [Thu, 30 Oct 2003 00:55:51 +0000 (00:55 +0000)] 
do not attempt to checksum if the LSP has been purged

21 years agoupdate from [email protected]
itojun [Wed, 29 Oct 2003 03:54:14 +0000 (03:54 +0000)] 
update from [email protected]

* RFC3315: DHCPv6
* supported DHCPv6 options:
*  RFC3319,
*  draft-ietf-dhc-dhcpv6-opt-dnsconfig-04.txt,
*  draft-ietf-dhc-dhcpv6-opt-prefix-delegation-05.txt
*  draft-ietf-dhc-dhcpv6-opt-timeconfig-02.txt,

21 years agomy address list option, from [email protected]
itojun [Wed, 29 Oct 2003 02:07:57 +0000 (02:07 +0000)] 
my address list option, from [email protected]

21 years agomisc. cleanup & display cosmetics [last for today i promise ;-)]
hannes [Tue, 28 Oct 2003 20:52:25 +0000 (20:52 +0000)] 
misc. cleanup & display cosmetics [last for today i promise ;-)]

21 years agodisplay the DIS lan-id of an IIH in non-verbose mode
hannes [Tue, 28 Oct 2003 19:10:15 +0000 (19:10 +0000)] 
display the DIS lan-id of an IIH in non-verbose mode

21 years agobackout 1.2
itojun [Tue, 28 Oct 2003 03:18:00 +0000 (03:18 +0000)] 
backout 1.2

21 years agounsigned/signed mixup
itojun [Tue, 28 Oct 2003 03:16:35 +0000 (03:16 +0000)] 
unsigned/signed mixup

21 years agostyle
itojun [Tue, 28 Oct 2003 03:14:16 +0000 (03:14 +0000)] 
style

21 years ago- move UDP port numbers to udp.h
hannes [Mon, 27 Oct 2003 22:44:37 +0000 (22:44 +0000)] 
- move UDP port numbers to udp.h
  - then the number defines# are acessible from outside

- make a differentiation between BFD_ECHO_PORT and BFD_CONTROL_PORT

21 years agodisplay cosmetics: stay inline to other decoders for vflag handling
hannes [Mon, 27 Oct 2003 17:21:32 +0000 (17:21 +0000)] 
display cosmetics: stay inline to other decoders for vflag handling

21 years agosupport for bi-directional fault detection (BFD) Control packets
hannes [Mon, 27 Oct 2003 10:13:44 +0000 (10:13 +0000)] 
support for bi-directional fault detection (BFD) Control packets
  as per draft-katz-ward-bfd-01.txt

21 years agofix for last commit: pass the correct pointer for the route target buffer
hannes [Mon, 27 Oct 2003 08:18:57 +0000 (08:18 +0000)] 
fix for last commit: pass the correct pointer for the route target buffer

21 years ago- support for the route target information SAFI #132 defined in draft-marques-ppvpn...
hannes [Mon, 27 Oct 2003 08:04:52 +0000 (08:04 +0000)] 
- support for the route target information SAFI #132 defined in draft-marques-ppvpn-rt-constrain-01.txt
- support for 32-bit AS number route-target (type 2) format

21 years agodisplay SNPAs rather than IS-Neighbors in TLV #6, addition of Flag field for the...
hannes [Sun, 26 Oct 2003 09:59:11 +0000 (09:59 +0000)] 
display SNPAs rather than IS-Neighbors in TLV #6, addition of Flag field for the shared risk link group GMPLS TLV

21 years agodisplay cosmetics: align the p2p IIH output to the LAN IIH output
hannes [Sat, 25 Oct 2003 12:02:01 +0000 (12:02 +0000)] 
display cosmetics: align the p2p IIH output to the LAN IIH output

21 years agoprint PDU length for PSNPs
hannes [Wed, 22 Oct 2003 22:32:01 +0000 (22:32 +0000)] 
print PDU length for PSNPs

21 years agoSquelch some compiler sarnings.
guy [Wed, 22 Oct 2003 20:00:03 +0000 (20:00 +0000)] 
Squelch some compiler sarnings.

Add some length checks.

21 years ago- fix a compiler padding issue in the LS-Request structure
hannes [Wed, 22 Oct 2003 17:08:45 +0000 (17:08 +0000)] 
- fix a compiler padding issue in the LS-Request structure
    u_int32_t ls_type becomes
    u_int8_t  ls_type[4];

- teach the LS-Request decoder howto properly print Opaque LSAs

21 years agomisc. cosmetic output & identation
hannes [Wed, 22 Oct 2003 16:29:18 +0000 (16:29 +0000)] 
misc. cosmetic output & identation

21 years ago- port LS_TYPE_OPAQUE printing from ospf_print_lsa() to ospf_print_lshdr()
hannes [Wed, 22 Oct 2003 15:47:44 +0000 (15:47 +0000)] 
- port LS_TYPE_OPAQUE printing from ospf_print_lsa() to ospf_print_lshdr()
  -that fixes the LSA-ACK printing of Opaque LSAs
   bugreport courtesy Carles Kishimoto Bisbe ckishimo [AT] ac [DOT] upc [DOT] es
- make use of ospf_print_lshdr() inside ospf_print_lsa()
- shared link risk group decoder from Carles Kishimoto Bisbe ckishimo [AT] ac [DOT] upc [DOT] es
- print IP adresses for LINK_IDs

21 years ago-add support for decoding the RSVP header Flag field
hannes [Wed, 22 Oct 2003 11:43:04 +0000 (11:43 +0000)] 
-add support for decoding the RSVP header Flag field
-add rfc2961 "Refresh reduction" Header Flag value

21 years agomisc. cosmetic changes to stay compliant to other multi-line decoders
hannes [Tue, 21 Oct 2003 23:04:23 +0000 (23:04 +0000)] 
misc. cosmetic changes to stay compliant to other multi-line decoders

21 years agomisc. cosmetic changes to stay compliant to other multi-line decoders
hannes [Tue, 21 Oct 2003 23:02:41 +0000 (23:02 +0000)] 
misc. cosmetic changes to stay compliant to other multi-line decoders

21 years agobug report [plus partial patch] from Carles Kishimoto Bisbe ckishimo [AT] ac [DOT...
hannes [Mon, 20 Oct 2003 16:11:45 +0000 (16:11 +0000)] 
bug report [plus partial patch] from Carles Kishimoto Bisbe ckishimo [AT] ac [DOT] upc [DOT] es
  -parse TE-TLVs until the ls_length field is zero
  -print router-ID TLV

21 years agodo a hexdump for the RSVP_HOP object for the C_Types which
hannes [Mon, 20 Oct 2003 10:50:09 +0000 (10:50 +0000)] 
do a hexdump for the RSVP_HOP object for the C_Types which
  do contain TLVs as we do not yet have a TLV decoder here;
print the codepoint for GMPLS LSP encodings

21 years agodisplay cosmetics: add missing colons and show codepoints for gmpls
hannes [Mon, 20 Oct 2003 10:35:20 +0000 (10:35 +0000)] 
display cosmetics: add missing colons and show codepoints for gmpls

21 years agoadd GMPLS payload ID codepoints mentioned in draft-ietf-ccamp-gmpls-g709-04.txt
hannes [Mon, 20 Oct 2003 10:32:46 +0000 (10:32 +0000)] 
add GMPLS payload ID codepoints mentioned in draft-ietf-ccamp-gmpls-g709-04.txt

21 years agoprint opcode number always
hannes [Mon, 20 Oct 2003 08:26:49 +0000 (08:26 +0000)] 
print opcode number always

21 years agoFrom Krzysztof Halasa: ignore single octet IEs.
guy [Sat, 18 Oct 2003 00:27:16 +0000 (00:27 +0000)] 
From Krzysztof Halasa: ignore single octet IEs.

21 years agoFrom Krzysztof Halasa <[email protected]>:
guy [Thu, 16 Oct 2003 09:50:36 +0000 (09:50 +0000)] 
From Krzysztof Halasa <[email protected]>:

1. Unused things (mbuf, rtentry) have been removed.

2. I've dropped FR_CR_BIT etc. definitions as they have only meaning in
   specific location in Q.922 address. FR_EA_BIT is still there as it's
   used by all Q.922 bytes.

3. a new parse_q922_addr() parses the whole Q.922 address structure
   and produces DLCI, Q.922 byte count and flags in ASCII. While I can't
   test 3- and 4-byte Q.922 addresses (using only default 2-byte ones),
   they are supported.

4. I've trimmed comments regarding HDLC flags and CRC/FCS bytes in FR
   packet. They are wire-only things and as such are never seen by
   libpcap/tcpdump.

5. Changed NLPID_LMI into NLPID_CISCO_LMI and NLPID_Q933 into NLPID_LMI
   (the latter is used by both ANSI (T1.617 annex D) and CCITT (Q.933
   annex A) LMI).

6. fr_hdrlen() now correctly returns FR header length (4 - 6 bytes,
   7 bytes with 4-byte Q.922 and incorrectly used pad byte).

7. I've changed output text formatting: "xxx-value-yyy-value" into
   "xxx value, yyy value", to be consistent with Ethernet.
   Not sure if it's correct.

8. added IPv6 support

9. q933_print() now reads lmi_print()

10. CCITT (Q.933 annex A) LMI is now supported, and both CCITT and ANSI
    fields are now correctly displayed

10. lmi_print() no longer prints constant fields (such as always zeroed
    Q.922 C/R bit or LMI Call Reference byte) unless they are set
    incorrectly.

11. I've added ANSI_ and CCITT_ prefixes to appropriate #defines.

12. a new parse_dlci_el() should handle all (10, 16 and 23-bit)
    DLCIs correctly (corresponding to 2, 3 and 4-byte Q.922 addresses,
    respectively).

21 years ago- squelch some compiler warnings
hannes [Sat, 4 Oct 2003 14:29:52 +0000 (14:29 +0000)] 
- squelch some compiler warnings
    courtesy Neil Spring nspring [AT] cs [DOT] washington [DOT] edu
- add support for the Traffic Engineering TLV plus all known subTLVs
- add support for draft-ietf-ccamp-ospf-gmpls-extensions defined subTLVs

21 years agoremove some LSA header code redundancy
hannes [Fri, 3 Oct 2003 13:20:46 +0000 (13:20 +0000)] 
remove some LSA header code redundancy

21 years agoshow opaque-LSA specific fields of the LSA header
hannes [Fri, 3 Oct 2003 13:00:33 +0000 (13:00 +0000)] 
show opaque-LSA specific fields of the LSA header

21 years agoadd better handling for unknown LSAs to OSPF
hannes [Thu, 2 Oct 2003 13:35:52 +0000 (13:35 +0000)] 
add better handling for unknown LSAs to OSPF

21 years agofix display of correct offset for hexdumps
hannes [Thu, 2 Oct 2003 13:16:14 +0000 (13:16 +0000)] 
fix display of correct offset for hexdumps

21 years agoFrom Scott Rose <[email protected]>:
guy [Thu, 25 Sep 2003 22:30:22 +0000 (22:30 +0000)] 
From Scott Rose <[email protected]>:

The test for DNS message dnssec bit flags in the query and
response are backwards.  The CD bit has meaning in a query, and
the AD bit has meaning in a response.

21 years agoAdd to the packet counts a count of packets that tcpdump has seen and
guy [Tue, 16 Sep 2003 21:02:51 +0000 (21:02 +0000)] 
Add to the packet counts a count of packets that tcpdump has seen and
processed - people often get confused by the "received by filter" count,
as they might not realize that, in some systems, it counts packets that
were handed to the filter mechanism regardless of whether they pass or,
if they pass, they have yet been read by the application and, on some
other systems, it counts packets that passed the filter but still counts
them in the kernel before they're read by the application.

They're probably mostly interested in the count of packets processed by
tcpdump; the "received by filter" count is probably of interest only if
you want to know how effective your filter is (and even then it's of
interest only if it counts packets handed to the filter rather than
packets that passed the filter).

Give more details on why the "received by filter" count doesn't
necessarily mean what you think it means.

21 years agoThe now-expired draft-perkins-aodv6-01 specified a different order for
guy [Sat, 13 Sep 2003 01:34:42 +0000 (01:34 +0000)] 
The now-expired draft-perkins-aodv6-01 specified a different order for
fields in most of the AODV messages when used with V6; print those
correctly.

Get rid of some "v6" tags that weren't there before my previous checkin.

21 years agoUse %[l]u, not %[l]d, to print unsigned quantities.
guy [Fri, 12 Sep 2003 22:16:53 +0000 (22:16 +0000)] 
Use %[l]u, not %[l]d, to print unsigned quantities.

21 years agoAs itojun said, u_int32_t can be long or int, depending on the platform.
guy [Fri, 12 Sep 2003 22:10:42 +0000 (22:10 +0000)] 
As itojun said, u_int32_t can be long or int, depending on the platform.

21 years agoUpdates from Bruce Simpson:
guy [Fri, 12 Sep 2003 22:05:56 +0000 (22:05 +0000)] 
Updates from Bruce Simpson:

Fix numerous bugs in my AODV submission, red faces all round,
mainly due to misunderstanding of EXTRACT_STRING() macro
arguments (because these are macros, type checking didn't
happen).

AODV HELLO extensions are now correctly printed.

The code is closer to FreeBSD's style guidelines.

Don't use "u_char" and "u_long", use "u_int8_t" and "u_int32_t".
"u_long" is definitely wrong as it's 64 bits on some platforms; "u_char"
is safe, but the official tcpdump style uses "u_int8_t".

At least according to the (long-since-expired) draft-perkins-aodv6-01,
the V6 AODV messages have different message type values from the V4 ones
in draft-ietf-manet-aodv-13.  However, draft-ietf-manet-aodv-13 says
that the only differences are in the addresses; we support both.

Clean up the way we tell the AODV6 dissector that we're running atop
IPv6 a bit.

Use "udpipaddr_print()" if we're printing AODV packets as a result of
saying "anything that matches this filter is AODV", rather than doing it
ourselves under the assumption that it's running over IPv4.

21 years agoAs noted in a bug report, the type for a topology change notification is
guy [Fri, 12 Sep 2003 20:53:19 +0000 (20:53 +0000)] 
As noted in a bug report, the type for a topology change notification is
0x80, not 1.

21 years agoupdate some references to internet drafts and change subTLV 4 to reflect the changes...
hannes [Mon, 8 Sep 2003 16:18:06 +0000 (16:18 +0000)] 
update some references to internet drafts and change subTLV 4 to reflect the changes of draft-ietf-isis-gmpls-extensions being now a local/remote ID

21 years agodo not forget to increment the pointer while reading pre-emption priorities
hannes [Mon, 8 Sep 2003 15:59:08 +0000 (15:59 +0000)] 
do not forget to increment the pointer while reading pre-emption priorities

21 years agoSquelch a compiler warning.
guy [Sat, 6 Sep 2003 22:16:56 +0000 (22:16 +0000)] 
Squelch a compiler warning.

21 years agominor display cosmetics to get to an consistent output style
hannes [Fri, 5 Sep 2003 17:42:41 +0000 (17:42 +0000)] 
minor display cosmetics to get to an consistent output style

21 years agouse strlcpy/cat
itojun [Wed, 13 Aug 2003 02:28:21 +0000 (02:28 +0000)] 
use strlcpy/cat

21 years agouse snprintf
itojun [Wed, 13 Aug 2003 02:26:52 +0000 (02:26 +0000)] 
use snprintf

21 years agou_int32_t is long or int depending on platform
itojun [Wed, 13 Aug 2003 02:23:07 +0000 (02:23 +0000)] 
u_int32_t is long or int depending on platform

21 years agoFrom Mark Ellzey Thomas <[email protected]>: some additional
guy [Sun, 10 Aug 2003 19:47:39 +0000 (19:47 +0000)] 
From Mark Ellzey Thomas <[email protected]>: some additional
buffer-overflow checking.

21 years agoCleanup of the Windows portion. The result of the GetVersion() function, that obtains...
risso [Fri, 8 Aug 2003 09:47:45 +0000 (09:47 +0000)] 
Cleanup of the Windows portion. The result of the GetVersion() function, that obtains the Windows OS version, was never used: avoid to call it.
Removed the call to PrinCapBegins in w32_fzs.c.

21 years agoRemoved w32_fzs.c from the Windows project
risso [Fri, 8 Aug 2003 09:42:23 +0000 (09:42 +0000)] 
Removed w32_fzs.c from the Windows project

21 years agoRemoved w32_fzs.c, since the functions it contains are no more needed by tcpdump
risso [Fri, 8 Aug 2003 09:41:09 +0000 (09:41 +0000)] 
Removed w32_fzs.c, since the functions it contains are no more needed by tcpdump

21 years agoModified the included files in order to make print-aodv compile under Windows as...
risso [Thu, 7 Aug 2003 11:31:42 +0000 (11:31 +0000)] 
Modified the included files in order to make print-aodv compile under Windows as well.

21 years agoAdded the new gmpls.c and print-aodv.c files to the Win32 project.
risso [Thu, 7 Aug 2003 11:29:30 +0000 (11:29 +0000)] 
Added the new gmpls.c and print-aodv.c files to the Win32 project.

21 years agoFrom Bruce M. Simpson <[email protected]>: RFC 3561 AODV support.
guy [Wed, 6 Aug 2003 06:49:38 +0000 (06:49 +0000)] 
From Bruce M. Simpson <[email protected]>: RFC 3561 AODV support.