1.\" @(#) $Header: /tcpdump/master/tcpdump/tcpdump.1.in,v 1.2 2008-11-09 23:35:03 mcr Exp $ (LBL)
2.\"
3.\" $NetBSD: tcpdump.8,v 1.9 2003/03/31 00:18:17 perry Exp $
4.\"
5.\" Copyright (c) 1987, 1988, 1989, 1990, 1991, 1992, 1994, 1995, 1996, 1997
6.\" The Regents of the University of California. All rights reserved.
7.\" All rights reserved.
8.\"
9.\" Redistribution and use in source and binary forms, with or without
10.\" modification, are permitted provided that: (1) source code distributions
11.\" retain the above copyright notice and this paragraph in its entirety, (2)
12.\" distributions including binary code include the above copyright notice and
13.\" this paragraph in its entirety in the documentation or other materials
14.\" provided with the distribution, and (3) all advertising materials mentioning
15.\" features or use of this software display the following acknowledgement:
16.\" ``This product includes software developed by the University of California,
17.\" Lawrence Berkeley Laboratory and its contributors.'' Neither the name of
18.\" the University nor the names of its contributors may be used to endorse
19.\" or promote products derived from this software without specific prior
20.\" written permission.
21.\" THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED
22.\" WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
23.\" MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
24.\"
25.TH TCPDUMP 1 "05 March 2009"
26.SH NAME
27tcpdump \- dump traffic on a network
28.SH SYNOPSIS
29.na
30.B tcpdump
31[
32.B \-AbdDefhHIJKlLnNOpqRStuUvxX
33] [
34.B \-B
35.I buffer_size
36] [
37.B \-c
38.I count
39]
40.br
41.ti +8
42[
43.B \-C
44.I file_size
45] [
46.B \-G
47.I rotate_seconds
48] [
49.B \-F
50.I file
51]
52.br
53.ti +8
54[
55.B \-i
56.I interface
57]
58[
59.B \-j
60.I tstamp_type
61]
62[
63.B \-m
64.I module
65]
66[
67.B \-M
68.I secret
69]
70.br
71.ti +8
72[
73.B \-r
74.I file
75]
76[
77.B \-s
78.I snaplen
79]
80[
81.B \-T
82.I type
83]
84[
85.B \-w
86.I file
87]
88.br
89.ti +8
90[
91.B \-W
92.I filecount
93]
94.br
95.ti +8
96[
97.B \-E
98.I spi@ipaddr algo:secret,...
99]
100.br
101.ti +8
102[
103.B \-y
104.I datalinktype
105]
106[
107.B \-z
108.I postrotate-command
109]
110[
111.B \-Z
112.I user
113]
114.ti +8
115[
116.I expression
117]
118.br
119.ad
120.SH DESCRIPTION
121.LP
122\fITcpdump\fP prints out a description of the contents of packets on a
123network interface that match the boolean \fIexpression\fP. It can also
124be run with the
125.B \-w
126flag, which causes it to save the packet data to a file for later
127analysis, and/or with the
128.B \-r
129flag, which causes it to read from a saved packet file rather than to
130read packets from a network interface. In all cases, only packets that
131match
132.I expression
133will be processed by
134.IR tcpdump .
135.LP
136.I Tcpdump
137will, if not run with the
138.B \-c
139flag, continue capturing packets until it is interrupted by a SIGINT
140signal (generated, for example, by typing your interrupt character,
141typically control-C) or a SIGTERM signal (typically generated with the
142.BR kill (1)
143command); if run with the
144.B \-c
145flag, it will capture packets until it is interrupted by a SIGINT or
146SIGTERM signal or the specified number of packets have been processed.
147.LP
148When
149.I tcpdump
150finishes capturing packets, it will report counts of:
151.IP
152packets ``captured'' (this is the number of packets that
153.I tcpdump
154has received and processed);
155.IP
156packets ``received by filter'' (the meaning of this depends on the OS on
157which you're running
158.IR tcpdump ,
159and possibly on the way the OS was configured - if a filter was
160specified on the command line, on some OSes it counts packets regardless
161of whether they were matched by the filter expression and, even if they
162were matched by the filter expression, regardless of whether
163.I tcpdump
164has read and processed them yet, on other OSes it counts only packets that were
165matched by the filter expression regardless of whether
166.I tcpdump
167has read and processed them yet, and on other OSes it counts only
168packets that were matched by the filter expression and were processed by
169.IR tcpdump );
170.IP
171packets ``dropped by kernel'' (this is the number of packets that were
172dropped, due to a lack of buffer space, by the packet capture mechanism
173in the OS on which
174.I tcpdump
175is running, if the OS reports that information to applications; if not,
176it will be reported as 0).
177.LP
178On platforms that support the SIGINFO signal, such as most BSDs
179(including Mac OS X) and Digital/Tru64 UNIX, it will report those counts
180when it receives a SIGINFO signal (generated, for example, by typing
181your ``status'' character, typically control-T, although on some
182platforms, such as Mac OS X, the ``status'' character is not set by
183default, so you must set it with
184.BR stty (1)
185in order to use it) and will continue capturing packets.
186.LP
187Reading packets from a network interface may require that you have
188special privileges; see the
189.B pcap (3PCAP)
190man page for details. Reading a saved packet file doesn't require
191special privileges.
192.SH OPTIONS
193.TP
194.B \-A
195Print each packet (minus its link level header) in ASCII. Handy for
196capturing web pages.
197.TP
198.B \-b
199Print the AS number in BGP packets in ASDOT notation rather than ASPLAIN
200notation.
201.TP
202.B \-B
|
205.TP
206.B \-c
207Exit after receiving \fIcount\fP packets.
208.TP
209.B \-C
210Before writing a raw packet to a savefile, check whether the file is
211currently larger than \fIfile_size\fP and, if so, close the current
212savefile and open a new one. Savefiles after the first savefile will
213have the name specified with the
214.B \-w
215flag, with a number after it, starting at 1 and continuing upward.
216The units of \fIfile_size\fP are millions of bytes (1,000,000 bytes,
217not 1,048,576 bytes).
218.TP
219.B \-d
220Dump the compiled packet-matching code in a human readable form to
221standard output and stop.
222.TP
223.B \-dd
224Dump packet-matching code as a
225.B C
226program fragment.
227.TP
228.B \-ddd
229Dump packet-matching code as decimal numbers (preceded with a count).
230.TP
231.B \-D
232Print the list of the network interfaces available on the system and on
233which
234.I tcpdump
235can capture packets. For each network interface, a number and an
236interface name, possibly followed by a text description of the
237interface, is printed. The interface name or the number can be supplied
238to the
239.B \-i
240flag to specify an interface on which to capture.
241.IP
242This can be useful on systems that don't have a command to list them
243(e.g., Windows systems, or UNIX systems lacking
244.BR "ifconfig \-a" );
245the number can be useful on Windows 2000 and later systems, where the
246interface name is a somewhat complex string.
247.IP
248The
249.B \-D
250flag will not be supported if
251.I tcpdump
252was built with an older version of
253.I libpcap
254that lacks the
255.B pcap_findalldevs()
256function.
257.TP
258.B \-e
259Print the link-level header on each dump line.
260.TP
261.B \-E
262Use \fIspi@ipaddr algo:secret\fP for decrypting IPsec ESP packets that
263are addressed to \fIaddr\fP and contain Security Parameter Index value
264\fIspi\fP. This combination may be repeated with comma or newline separation.
265.IP
266Note that setting the secret for IPv4 ESP packets is supported at this time.
267.IP
268Algorithms may be
269\fBdes-cbc\fP,
270\fB3des-cbc\fP,
271\fBblowfish-cbc\fP,
272\fBrc3-cbc\fP,
273\fBcast128-cbc\fP, or
274\fBnone\fP.
275The default is \fBdes-cbc\fP.
276The ability to decrypt packets is only present if \fItcpdump\fP was compiled
277with cryptography enabled.
278.IP
279\fIsecret\fP is the ASCII text for ESP secret key.
280If preceded by 0x, then a hex value will be read.
281.IP
282The option assumes RFC2406 ESP, not RFC1827 ESP.
283The option is only for debugging purposes, and
284the use of this option with a true `secret' key is discouraged.
285By presenting IPsec secret key onto command line
286you make it visible to others, via
287.IR ps (1)
288and other occasions.
289.IP
290In addition to the above syntax, the syntax \fIfile name\fP may be used
291to have tcpdump read the provided file in. The file is opened upon
292receiving the first ESP packet, so any special permissions that tcpdump
293may have been given should already have been given up.
294.TP
295.B \-f
296Print `foreign' IPv4 addresses numerically rather than symbolically
297(this option is intended to get around serious brain damage in
298Sun's NIS server \(em usually it hangs forever translating non-local
299internet numbers).
300.IP
301The test for `foreign' IPv4 addresses is done using the IPv4 address and
302netmask of the interface on which capture is being done. If that
303address or netmask are not available, available, either because the
304interface on which capture is being done has no address or netmask or
305because the capture is being done on the Linux "any" interface, which
306can capture on more than one interface, this option will not work
307correctly.
308.TP
309.B \-F
310Use \fIfile\fP as input for the filter expression.
311An additional expression given on the command line is ignored.
312.TP
313.B \-G
314If specified, rotates the dump file specified with the
315.B \-w
316option every \fIrotate_seconds\fP seconds.
317Savefiles will have the name specified by
318.B \-w
319which should include a time format as defined by
320.BR strftime (3).
321If no time format is specified, each new file will overwrite the previous.
322.IP
323If used in conjunction with the
324.B \-C
325option, filenames will take the form of `\fIfile\fP<count>'.
326.TP
327.B \-h
328Print the tcpdump and libpcap version strings, print a usage message,
329and exit.
330.TP
331.B \-H
332Attempt to detect 802.11s draft mesh headers.
333.TP
334.B \-i
335Listen on \fIinterface\fP.
336If unspecified, \fItcpdump\fP searches the system interface list for the
337lowest numbered, configured up interface (excluding loopback).
338Ties are broken by choosing the earliest match.
339.IP
340On Linux systems with 2.2 or later kernels, an
341.I interface
342argument of ``any'' can be used to capture packets from all interfaces.
343Note that captures on the ``any'' device will not be done in promiscuous
344mode.
345.IP
346If the
347.B \-D
348flag is supported, an interface number as printed by that flag can be
349used as the
350.I interface
351argument.
352.TP
353.B \-I
354Put the interface in "monitor mode"; this is supported only on IEEE
355802.11 Wi-Fi interfaces, and supported only on some operating systems.
356.IP
357Note that in monitor mode the adapter might disassociate from the
358network with which it's associated, so that you will not be able to use
359any wireless networks with that adapter. This could prevent accessing
360files on a network server, or resolving host names or network addresses,
361if you are capturing in monitor mode and are not connected to another
362network with another adapter.
363.IP
364This flag will affect the output of the
365.B \-L
366flag. If
367.B \-I
368isn't specified, only those link-layer types available when not in
369monitor mode will be shown; if
370.B \-I
371is specified, only those link-layer types available when in monitor mode
372will be shown.
373.TP
374.B \-j
375Set the time stamp type for the capture to \fItstamp_type\fP. The names
376to use for the time stamp types are given in
377.BR pcap-tstamp-type (@MAN_MISC_INFO@);
378not all the types listed there will necessarily be valid for any given
379interface.
380.TP
381.B \-J
382List the supported time stamp types for the interface and exit. If the
383time stamp type cannot be set for the interface, no time stamp types are
384listed.
385.TP
386.B \-K
387Don't attempt to verify IP, TCP, or UDP checksums. This is useful for
388interfaces that perform some or all of those checksum calculation in
389hardware; otherwise, all outgoing TCP checksums will be flagged as bad.
390.TP
391.B \-l
392Make stdout line buffered.
393Useful if you want to see the data
394while capturing it.
395E.g.,
|
605See
606.BR pcap-savefile (@MAN_FILE_FORMATS@)
607for a description of the file format.
608.TP
609.B \-W
610Used in conjunction with the
611.B \-C
612option, this will limit the number
613of files created to the specified number, and begin overwriting files
614from the beginning, thus creating a 'rotating' buffer.
615In addition, it will name
616the files with enough leading 0s to support the maximum number of
617files, allowing them to sort correctly.
618.IP
619Used in conjunction with the
620.B \-G
621option, this will limit the number of rotated dump files that get
622created, exiting with status 0 when reaching the limit. If used with
623.B \-C
624as well, the behavior will result in cyclical files per timeslice.
625.TP
626.B \-x
627When parsing and printing,
628in addition to printing the headers of each packet, print the data of
629each packet (minus its link level header) in hex.
630The smaller of the entire packet or
631.I snaplen
632bytes will be printed. Note that this is the entire link-layer
633packet, so for link layers that pad (e.g. Ethernet), the padding bytes
634will also be printed when the higher layer packet is shorter than the
635required padding.
636.TP
637.B \-xx
638When parsing and printing,
639in addition to printing the headers of each packet, print the data of
640each packet,
641.I including
642its link level header, in hex.
643.TP
644.B \-X
645When parsing and printing,
646in addition to printing the headers of each packet, print the data of
647each packet (minus its link level header) in hex and ASCII.
648This is very handy for analysing new protocols.
649.TP
650.B \-XX
651When parsing and printing,
652in addition to printing the headers of each packet, print the data of
653each packet,
654.I including
655its link level header, in hex and ASCII.
656.TP
657.B \-y
658Set the data link type to use while capturing packets to \fIdatalinktype\fP.
659.TP
660.B \-z
661Used in conjunction with the
662.B -C
663or
664.B -G
665options, this will make
666.I tcpdump
667run "
668.I command file
669" where
670.I file
671is the savefile being closed after each rotation. For example, specifying
672.B \-z gzip
673or
674.B \-z bzip2
675will compress each savefile using gzip or bzip2.
676.IP
677Note that tcpdump will run the command in parallel to the capture, using
678the lowest priority so that this doesn't disturb the capture process.
679.IP
680And in case you would like to use a command that itself takes flags or
681different arguments, you can always write a shell script that will take the
682savefile name as the only argument, make the flags & arguments arrangements
683and execute the command that you want.
684.TP
685.B \-Z
686If
687.I tcpdump
688is running as root, after opening the capture device or input savefile,
689but before opening any savefiles for output, change the user ID to
690.I user
691and the group ID to the primary group of
692.IR user .
693.IP
694This behavior can also be enabled by default at compile time.
695.IP "\fI expression\fP"
696.RS
697selects which packets will be dumped.
698If no \fIexpression\fP
699is given, all packets on the net will be dumped.
700Otherwise,
701only packets for which \fIexpression\fP is `true' will be dumped.
702.LP
703For the \fIexpression\fP syntax, see
704.BR pcap-filter (@MAN_MISC_INFO@).
705.LP
706Expression arguments can be passed to \fItcpdump\fP as either a single
707argument or as multiple arguments, whichever is more convenient.
708Generally, if the expression contains Shell metacharacters, it is
709easier to pass it as a single, quoted argument.
710Multiple arguments are concatenated with spaces before being parsed.
711.SH EXAMPLES
712.LP
713To print all packets arriving at or departing from \fIsundown\fP:
714.RS
715.nf
716\fBtcpdump host sundown\fP
717.fi
718.RE
719.LP
720To print traffic between \fIhelios\fR and either \fIhot\fR or \fIace\fR:
721.RS
722.nf
723\fBtcpdump host helios and \\( hot or ace \\)\fP
724.fi
725.RE
726.LP
727To print all IP packets between \fIace\fR and any host except \fIhelios\fR:
728.RS
729.nf
730\fBtcpdump ip host ace and not helios\fP
731.fi
732.RE
733.LP
734To print all traffic between local hosts and hosts at Berkeley:
735.RS
736.nf
737.B
738tcpdump net ucb-ether
739.fi
740.RE
741.LP
742To print all ftp traffic through internet gateway \fIsnup\fP:
743(note that the expression is quoted to prevent the shell from
744(mis-)interpreting the parentheses):
745.RS
746.nf
747.B
748tcpdump 'gateway snup and (port ftp or ftp-data)'
749.fi
750.RE
751.LP
752To print traffic neither sourced from nor destined for local hosts
753(if you gateway to one other net, this stuff should never make it
754onto your local net).
755.RS
756.nf
757.B
758tcpdump ip and not net \fIlocalnet\fP
759.fi
760.RE
761.LP
762To print the start and end packets (the SYN and FIN packets) of each
763TCP conversation that involves a non-local host.
764.RS
765.nf
766.B
767tcpdump 'tcp[tcpflags] & (tcp-syn|tcp-fin) != 0 and not src and dst net \fIlocalnet\fP'
768.fi
769.RE
770.LP
771To print all IPv4 HTTP packets to and from port 80, i.e. print only
772packets that contain data, not, for example, SYN and FIN packets and
773ACK-only packets. (IPv6 is left as an exercise for the reader.)
774.RS
775.nf
776.B
777tcpdump 'tcp port 80 and (((ip[2:2] - ((ip[0]&0xf)<<2)) - ((tcp[12]&0xf0)>>2)) != 0)'
778.fi
779.RE
780.LP
781To print IP packets longer than 576 bytes sent through gateway \fIsnup\fP:
782.RS
783.nf
784.B
785tcpdump 'gateway snup and ip[2:2] > 576'
786.fi
787.RE
788.LP
789To print IP broadcast or multicast packets that were
790.I not
791sent via Ethernet broadcast or multicast:
792.RS
793.nf
794.B
795tcpdump 'ether[0] & 1 = 0 and ip[16] >= 224'
796.fi
797.RE
798.LP
799To print all ICMP packets that are not echo requests/replies (i.e., not
800ping packets):
801.RS
802.nf
803.B
804tcpdump 'icmp[icmptype] != icmp-echo and icmp[icmptype] != icmp-echoreply'
805.fi
806.RE
807.SH OUTPUT FORMAT
808.LP
809The output of \fItcpdump\fP is protocol dependent.
810The following
811gives a brief description and examples of most of the formats.
812.de HD
813.sp 1.5
814.B
815..
816.HD
817Link Level Headers
818.LP
819If the '-e' option is given, the link level header is printed out.
820On Ethernets, the source and destination addresses, protocol,
821and packet length are printed.
822.LP
823On FDDI networks, the '-e' option causes \fItcpdump\fP to print
824the `frame control' field, the source and destination addresses,
825and the packet length.
826(The `frame control' field governs the
827interpretation of the rest of the packet.
828Normal packets (such
829as those containing IP datagrams) are `async' packets, with a priority
830value between 0 and 7; for example, `\fBasync4\fR'.
831Such packets
832are assumed to contain an 802.2 Logical Link Control (LLC) packet;
833the LLC header is printed if it is \fInot\fR an ISO datagram or a
834so-called SNAP packet.
835.LP
836On Token Ring networks, the '-e' option causes \fItcpdump\fP to print
837the `access control' and `frame control' fields, the source and
838destination addresses, and the packet length.
839As on FDDI networks,
840packets are assumed to contain an LLC packet.
841Regardless of whether
842the '-e' option is specified or not, the source routing information is
843printed for source-routed packets.
844.LP
845On 802.11 networks, the '-e' option causes \fItcpdump\fP to print
846the `frame control' fields, all of the addresses in the 802.11 header,
847and the packet length.
848As on FDDI networks,
849packets are assumed to contain an LLC packet.
850.LP
851\fI(N.B.: The following description assumes familiarity with
852the SLIP compression algorithm described in RFC-1144.)\fP
853.LP
854On SLIP links, a direction indicator (``I'' for inbound, ``O'' for outbound),
855packet type, and compression information are printed out.
856The packet type is printed first.
857The three types are \fIip\fP, \fIutcp\fP, and \fIctcp\fP.
858No further link information is printed for \fIip\fR packets.
859For TCP packets, the connection identifier is printed following the type.
860If the packet is compressed, its encoded header is printed out.
861The special cases are printed out as
862\fB*S+\fIn\fR and \fB*SA+\fIn\fR, where \fIn\fR is the amount by which
863the sequence number (or sequence number and ack) has changed.
864If it is not a special case,
865zero or more changes are printed.
866A change is indicated by U (urgent pointer), W (window), A (ack),
867S (sequence number), and I (packet ID), followed by a delta (+n or -n),
868or a new value (=n).
869Finally, the amount of data in the packet and compressed header length
870are printed.
871.LP
872For example, the following line shows an outbound compressed TCP packet,
873with an implicit connection identifier; the ack has changed by 6,
874the sequence number by 49, and the packet ID by 6; there are 3 bytes of
875data and 6 bytes of compressed header:
876.RS
877.nf
878\fBO ctcp * A+6 S+49 I+6 3 (6)\fP
879.fi
880.RE
881.HD
882ARP/RARP Packets
883.LP
884Arp/rarp output shows the type of request and its arguments.
885The
886format is intended to be self explanatory.
887Here is a short sample taken from the start of an `rlogin' from
888host \fIrtsg\fP to host \fIcsam\fP:
889.RS
890.nf
891.sp .5
892\f(CWarp who-has csam tell rtsg
893arp reply csam is-at CSAM\fR
894.sp .5
895.fi
896.RE
897The first line says that rtsg sent an arp packet asking
898for the Ethernet address of internet host csam.
899Csam
900replies with its Ethernet address (in this example, Ethernet addresses
901are in caps and internet addresses in lower case).
902.LP
903This would look less redundant if we had done \fItcpdump \-n\fP:
904.RS
905.nf
906.sp .5
907\f(CWarp who-has 128.3.254.6 tell 128.3.254.68
908arp reply 128.3.254.6 is-at 02:07:01:00:01:c4\fP
909.fi
910.RE
911.LP
912If we had done \fItcpdump \-e\fP, the fact that the first packet is
913broadcast and the second is point-to-point would be visible:
914.RS
915.nf
916.sp .5
917\f(CWRTSG Broadcast 0806 64: arp who-has csam tell rtsg
918CSAM RTSG 0806 64: arp reply csam is-at CSAM\fR
919.sp .5
920.fi
921.RE
922For the first packet this says the Ethernet source address is RTSG, the
923destination is the Ethernet broadcast address, the type field
924contained hex 0806 (type ETHER_ARP) and the total length was 64 bytes.
925.HD
926TCP Packets
927.LP
928\fI(N.B.:The following description assumes familiarity with
929the TCP protocol described in RFC-793.
930If you are not familiar
931with the protocol, neither this description nor \fItcpdump\fP will
932be of much use to you.)\fP
933.LP
934The general format of a tcp protocol line is:
935.RS
936.nf
937.sp .5
938\fIsrc > dst: flags data-seqno ack window urgent options\fP
939.sp .5
940.fi
941.RE
942\fISrc\fP and \fIdst\fP are the source and destination IP
943addresses and ports.
944\fIFlags\fP are some combination of S (SYN),
945F (FIN), P (PUSH), R (RST), U (URG), W (ECN CWR), E (ECN-Echo) or
946`.' (ACK), or `none' if no flags are set.
947\fIData-seqno\fP describes the portion of sequence space covered
948by the data in this packet (see example below).
949\fIAck\fP is sequence number of the next data expected the other
950direction on this connection.
951\fIWindow\fP is the number of bytes of receive buffer space available
952the other direction on this connection.
953\fIUrg\fP indicates there is `urgent' data in the packet.
954\fIOptions\fP are tcp options enclosed in angle brackets (e.g., <mss 1024>).
955.LP
956\fISrc, dst\fP and \fIflags\fP are always present.
957The other fields
958depend on the contents of the packet's tcp protocol header and
959are output only if appropriate.
960.LP
961Here is the opening portion of an rlogin from host \fIrtsg\fP to
962host \fIcsam\fP.
963.RS
964.nf
965.sp .5
966\s-2\f(CWrtsg.1023 > csam.login: S 768512:768512(0) win 4096 <mss 1024>
967csam.login > rtsg.1023: S 947648:947648(0) ack 768513 win 4096 <mss 1024>
968rtsg.1023 > csam.login: . ack 1 win 4096
969rtsg.1023 > csam.login: P 1:2(1) ack 1 win 4096
970csam.login > rtsg.1023: . ack 2 win 4096
971rtsg.1023 > csam.login: P 2:21(19) ack 1 win 4096
972csam.login > rtsg.1023: P 1:2(1) ack 21 win 4077
973csam.login > rtsg.1023: P 2:3(1) ack 21 win 4077 urg 1
974csam.login > rtsg.1023: P 3:4(1) ack 21 win 4077 urg 1\fR\s+2
975.sp .5
976.fi
977.RE
978The first line says that tcp port 1023 on rtsg sent a packet
979to port \fIlogin\fP
980on csam.
981The \fBS\fP indicates that the \fISYN\fP flag was set.
982The packet sequence number was 768512 and it contained no data.
983(The notation is `first:last(nbytes)' which means `sequence
984numbers \fIfirst\fP
985up to but not including \fIlast\fP which is \fInbytes\fP bytes of user data'.)
986There was no piggy-backed ack, the available receive window was 4096
987bytes and there was a max-segment-size option requesting an mss of
9881024 bytes.
989.LP
990Csam replies with a similar packet except it includes a piggy-backed
991ack for rtsg's SYN.
992Rtsg then acks csam's SYN.
993The `.' means the ACK flag was set.
994The packet contained no data so there is no data sequence number.
995Note that the ack sequence
996number is a small integer (1).
997The first time \fItcpdump\fP sees a
998tcp `conversation', it prints the sequence number from the packet.
999On subsequent packets of the conversation, the difference between
1000the current packet's sequence number and this initial sequence number
1001is printed.
1002This means that sequence numbers after the
1003first can be interpreted
1004as relative byte positions in the conversation's data stream (with the
1005first data byte each direction being `1').
1006`-S' will override this
1007feature, causing the original sequence numbers to be output.
1008.LP
1009On the 6th line, rtsg sends csam 19 bytes of data (bytes 2 through 20
1010in the rtsg \(-> csam side of the conversation).
1011The PUSH flag is set in the packet.
1012On the 7th line, csam says it's received data sent by rtsg up to
1013but not including byte 21.
1014Most of this data is apparently sitting in the
1015socket buffer since csam's receive window has gotten 19 bytes smaller.
1016Csam also sends one byte of data to rtsg in this packet.
1017On the 8th and 9th lines,
1018csam sends two bytes of urgent, pushed data to rtsg.
1019.LP
1020If the snapshot was small enough that \fItcpdump\fP didn't capture
1021the full TCP header, it interprets as much of the header as it can
1022and then reports ``[|\fItcp\fP]'' to indicate the remainder could not
1023be interpreted.
1024If the header contains a bogus option (one with a length
1025that's either too small or beyond the end of the header), \fItcpdump\fP
1026reports it as ``[\fIbad opt\fP]'' and does not interpret any further
1027options (since it's impossible to tell where they start).
1028If the header
1029length indicates options are present but the IP datagram length is not
1030long enough for the options to actually be there, \fItcpdump\fP reports
1031it as ``[\fIbad hdr length\fP]''.
1032.HD
1033.B Capturing TCP packets with particular flag combinations (SYN-ACK, URG-ACK, etc.)
1034.PP
1035There are 8 bits in the control bits section of the TCP header:
1036.IP
1037.I CWR | ECE | URG | ACK | PSH | RST | SYN | FIN
1038.PP
1039Let's assume that we want to watch packets used in establishing
1040a TCP connection.
1041Recall that TCP uses a 3-way handshake protocol
1042when it initializes a new connection; the connection sequence with
1043regard to the TCP control bits is
1044.PP
1045.RS
10461) Caller sends SYN
1047.RE
1048.RS
10492) Recipient responds with SYN, ACK
1050.RE
1051.RS
10523) Caller sends ACK
1053.RE
1054.PP
1055Now we're interested in capturing packets that have only the
1056SYN bit set (Step 1).
1057Note that we don't want packets from step 2
1058(SYN-ACK), just a plain initial SYN.
1059What we need is a correct filter
1060expression for \fItcpdump\fP.
1061.PP
1062Recall the structure of a TCP header without options:
1063.PP
1064.nf
1065 0 15 31
1066-----------------------------------------------------------------
1067| source port | destination port |
1068-----------------------------------------------------------------
1069| sequence number |
1070-----------------------------------------------------------------
1071| acknowledgment number |
1072-----------------------------------------------------------------
1073| HL | rsvd |C|E|U|A|P|R|S|F| window size |
1074-----------------------------------------------------------------
1075| TCP checksum | urgent pointer |
1076-----------------------------------------------------------------
1077.fi
1078.PP
1079A TCP header usually holds 20 octets of data, unless options are
1080present.
1081The first line of the graph contains octets 0 - 3, the
1082second line shows octets 4 - 7 etc.
1083.PP
1084Starting to count with 0, the relevant TCP control bits are contained
1085in octet 13:
1086.PP
1087.nf
1088 0 7| 15| 23| 31
1089----------------|---------------|---------------|----------------
1090| HL | rsvd |C|E|U|A|P|R|S|F| window size |
1091----------------|---------------|---------------|----------------
1092| | 13th octet | | |
1093.fi
1094.PP
1095Let's have a closer look at octet no. 13:
1096.PP
1097.nf
1098 | |
1099 |---------------|
1100 |C|E|U|A|P|R|S|F|
1101 |---------------|
1102 |7 5 3 0|
1103.fi
1104.PP
1105These are the TCP control bits we are interested
1106in.
1107We have numbered the bits in this octet from 0 to 7, right to
1108left, so the PSH bit is bit number 3, while the URG bit is number 5.
1109.PP
1110Recall that we want to capture packets with only SYN set.
1111Let's see what happens to octet 13 if a TCP datagram arrives
1112with the SYN bit set in its header:
1113.PP
1114.nf
1115 |C|E|U|A|P|R|S|F|
1116 |---------------|
1117 |0 0 0 0 0 0 1 0|
1118 |---------------|
1119 |7 6 5 4 3 2 1 0|
1120.fi
1121.PP
1122Looking at the
1123control bits section we see that only bit number 1 (SYN) is set.
1124.PP
1125Assuming that octet number 13 is an 8-bit unsigned integer in
1126network byte order, the binary value of this octet is
1127.IP
112800000010
1129.PP
1130and its decimal representation is
1131.PP
1132.nf
1133 7 6 5 4 3 2 1 0
11340*2 + 0*2 + 0*2 + 0*2 + 0*2 + 0*2 + 1*2 + 0*2 = 2
1135.fi
1136.PP
1137We're almost done, because now we know that if only SYN is set,
1138the value of the 13th octet in the TCP header, when interpreted
1139as a 8-bit unsigned integer in network byte order, must be exactly 2.
1140.PP
1141This relationship can be expressed as
1142.RS
1143.B
1144tcp[13] == 2
1145.RE
1146.PP
1147We can use this expression as the filter for \fItcpdump\fP in order
1148to watch packets which have only SYN set:
1149.RS
1150.B
1151tcpdump -i xl0 tcp[13] == 2
1152.RE
1153.PP
1154The expression says "let the 13th octet of a TCP datagram have
1155the decimal value 2", which is exactly what we want.
1156.PP
1157Now, let's assume that we need to capture SYN packets, but we
1158don't care if ACK or any other TCP control bit is set at the
1159same time.
1160Let's see what happens to octet 13 when a TCP datagram
1161with SYN-ACK set arrives:
1162.PP
1163.nf
1164 |C|E|U|A|P|R|S|F|
1165 |---------------|
1166 |0 0 0 1 0 0 1 0|
1167 |---------------|
1168 |7 6 5 4 3 2 1 0|
1169.fi
1170.PP
1171Now bits 1 and 4 are set in the 13th octet.
1172The binary value of
1173octet 13 is
1174.IP
1175 00010010
1176.PP
1177which translates to decimal
1178.PP
1179.nf
1180 7 6 5 4 3 2 1 0
11810*2 + 0*2 + 0*2 + 1*2 + 0*2 + 0*2 + 1*2 + 0*2 = 18
1182.fi
1183.PP
1184Now we can't just use 'tcp[13] == 18' in the \fItcpdump\fP filter
1185expression, because that would select only those packets that have
1186SYN-ACK set, but not those with only SYN set.
1187Remember that we don't care
1188if ACK or any other control bit is set as long as SYN is set.
1189.PP
1190In order to achieve our goal, we need to logically AND the
1191binary value of octet 13 with some other value to preserve
1192the SYN bit.
1193We know that we want SYN to be set in any case,
1194so we'll logically AND the value in the 13th octet with
1195the binary value of a SYN:
1196.PP
1197.nf
1198
1199 00010010 SYN-ACK 00000010 SYN
1200 AND 00000010 (we want SYN) AND 00000010 (we want SYN)
1201 -------- --------
1202 = 00000010 = 00000010
1203.fi
1204.PP
1205We see that this AND operation delivers the same result
1206regardless whether ACK or another TCP control bit is set.
1207The decimal representation of the AND value as well as
1208the result of this operation is 2 (binary 00000010),
1209so we know that for packets with SYN set the following
1210relation must hold true:
1211.IP
1212( ( value of octet 13 ) AND ( 2 ) ) == ( 2 )
1213.PP
1214This points us to the \fItcpdump\fP filter expression
1215.RS
1216.B
1217 tcpdump -i xl0 'tcp[13] & 2 == 2'
1218.RE
1219.PP
1220Some offsets and field values may be expressed as names
1221rather than as numeric values. For example tcp[13] may
1222be replaced with tcp[tcpflags]. The following TCP flag
1223field values are also available: tcp-fin, tcp-syn, tcp-rst,
1224tcp-push, tcp-act, tcp-urg.
1225.PP
1226This can be demonstrated as:
1227.RS
1228.B
1229 tcpdump -i xl0 'tcp[tcpflags] & tcp-push != 0'
1230.RE
1231.PP
1232Note that you should use single quotes or a backslash
1233in the expression to hide the AND ('&') special character
1234from the shell.
1235.HD
1236.B
1237UDP Packets
1238.LP
1239UDP format is illustrated by this rwho packet:
1240.RS
1241.nf
1242.sp .5
1243\f(CWactinide.who > broadcast.who: udp 84\fP
1244.sp .5
1245.fi
1246.RE
1247This says that port \fIwho\fP on host \fIactinide\fP sent a udp
1248datagram to port \fIwho\fP on host \fIbroadcast\fP, the Internet
1249broadcast address.
1250The packet contained 84 bytes of user data.
1251.LP
1252Some UDP services are recognized (from the source or destination
1253port number) and the higher level protocol information printed.
1254In particular, Domain Name service requests (RFC-1034/1035) and Sun
1255RPC calls (RFC-1050) to NFS.
1256.HD
1257UDP Name Server Requests
1258.LP
1259\fI(N.B.:The following description assumes familiarity with
1260the Domain Service protocol described in RFC-1035.
1261If you are not familiar
1262with the protocol, the following description will appear to be written
1263in greek.)\fP
1264.LP
1265Name server requests are formatted as
1266.RS
1267.nf
1268.sp .5
1269\fIsrc > dst: id op? flags qtype qclass name (len)\fP
1270.sp .5
1271\f(CWh2opolo.1538 > helios.domain: 3+ A? ucbvax.berkeley.edu. (37)\fR
1272.sp .5
1273.fi
1274.RE
1275Host \fIh2opolo\fP asked the domain server on \fIhelios\fP for an
1276address record (qtype=A) associated with the name \fIucbvax.berkeley.edu.\fP
1277The query id was `3'.
1278The `+' indicates the \fIrecursion desired\fP flag
1279was set.
1280The query length was 37 bytes, not including the UDP and
1281IP protocol headers.
1282The query operation was the normal one, \fIQuery\fP,
1283so the op field was omitted.
1284If the op had been anything else, it would
1285have been printed between the `3' and the `+'.
1286Similarly, the qclass was the normal one,
1287\fIC_IN\fP, and omitted.
1288Any other qclass would have been printed
1289immediately after the `A'.
1290.LP
1291A few anomalies are checked and may result in extra fields enclosed in
1292square brackets: If a query contains an answer, authority records or
1293additional records section,
1294.IR ancount ,
1295.IR nscount ,
1296or
1297.I arcount
1298are printed as `[\fIn\fPa]', `[\fIn\fPn]' or `[\fIn\fPau]' where \fIn\fP
1299is the appropriate count.
1300If any of the response bits are set (AA, RA or rcode) or any of the
1301`must be zero' bits are set in bytes two and three, `[b2&3=\fIx\fP]'
1302is printed, where \fIx\fP is the hex value of header bytes two and three.
1303.HD
1304UDP Name Server Responses
1305.LP
1306Name server responses are formatted as
1307.RS
1308.nf
1309.sp .5
1310\fIsrc > dst: id op rcode flags a/n/au type class data (len)\fP
1311.sp .5
1312\f(CWhelios.domain > h2opolo.1538: 3 3/3/7 A 128.32.137.3 (273)
1313helios.domain > h2opolo.1537: 2 NXDomain* 0/1/0 (97)\fR
1314.sp .5
1315.fi
1316.RE
1317In the first example, \fIhelios\fP responds to query id 3 from \fIh2opolo\fP
1318with 3 answer records, 3 name server records and 7 additional records.
1319The first answer record is type A (address) and its data is internet
1320address 128.32.137.3.
1321The total size of the response was 273 bytes,
1322excluding UDP and IP headers.
1323The op (Query) and response code
1324(NoError) were omitted, as was the class (C_IN) of the A record.
1325.LP
1326In the second example, \fIhelios\fP responds to query 2 with a
1327response code of non-existent domain (NXDomain) with no answers,
1328one name server and no authority records.
1329The `*' indicates that
1330the \fIauthoritative answer\fP bit was set.
1331Since there were no
1332answers, no type, class or data were printed.
1333.LP
1334Other flag characters that might appear are `\-' (recursion available,
1335RA, \fInot\fP set) and `|' (truncated message, TC, set).
1336If the
1337`question' section doesn't contain exactly one entry, `[\fIn\fPq]'
1338is printed.
1339.HD
1340SMB/CIFS decoding
1341.LP
1342\fItcpdump\fP now includes fairly extensive SMB/CIFS/NBT decoding for data
1343on UDP/137, UDP/138 and TCP/139.
1344Some primitive decoding of IPX and
1345NetBEUI SMB data is also done.
1346.LP
1347By default a fairly minimal decode is done, with a much more detailed
1348decode done if -v is used.
1349Be warned that with -v a single SMB packet
1350may take up a page or more, so only use -v if you really want all the
1351gory details.
1352.LP
1353For information on SMB packet formats and what all the fields mean see
1354www.cifs.org or the pub/samba/specs/ directory on your favorite
1355samba.org mirror site.
1356The SMB patches were written by Andrew Tridgell
1357(tridge@samba.org).
1358.HD
1359NFS Requests and Replies
1360.LP
1361Sun NFS (Network File System) requests and replies are printed as:
1362.RS
1363.nf
1364.sp .5
1365\fIsrc.xid > dst.nfs: len op args\fP
1366\fIsrc.nfs > dst.xid: reply stat len op results\fP
1367.sp .5
1368\f(CW
1369sushi.6709 > wrl.nfs: 112 readlink fh 21,24/10.73165
1370wrl.nfs > sushi.6709: reply ok 40 readlink "../var"
1371sushi.201b > wrl.nfs:
1372 144 lookup fh 9,74/4096.6878 "xcolors"
1373wrl.nfs > sushi.201b:
1374 reply ok 128 lookup fh 9,74/4134.3150
1375\fR
1376.sp .5
1377.fi
1378.RE
1379In the first line, host \fIsushi\fP sends a transaction with id \fI6709\fP
1380to \fIwrl\fP (note that the number following the src host is a
1381transaction id, \fInot\fP the source port).
1382The request was 112 bytes,
1383excluding the UDP and IP headers.
1384The operation was a \fIreadlink\fP
1385(read symbolic link) on file handle (\fIfh\fP) 21,24/10.731657119.
1386(If one is lucky, as in this case, the file handle can be interpreted
1387as a major,minor device number pair, followed by the inode number and
1388generation number.)
1389\fIWrl\fP replies `ok' with the contents of the link.
1390.LP
1391In the third line, \fIsushi\fP asks \fIwrl\fP to lookup the name
1392`\fIxcolors\fP' in directory file 9,74/4096.6878.
1393Note that the data printed
1394depends on the operation type.
1395The format is intended to be self
1396explanatory if read in conjunction with
1397an NFS protocol spec.
1398.LP
1399If the \-v (verbose) flag is given, additional information is printed.
1400For example:
1401.RS
1402.nf
1403.sp .5
1404\f(CW
1405sushi.1372a > wrl.nfs:
1406 148 read fh 21,11/12.195 8192 bytes @ 24576
1407wrl.nfs > sushi.1372a:
1408 reply ok 1472 read REG 100664 ids 417/0 sz 29388
1409\fP
1410.sp .5
1411.fi
1412.RE
1413(\-v also prints the IP header TTL, ID, length, and fragmentation fields,
1414which have been omitted from this example.) In the first line,
1415\fIsushi\fP asks \fIwrl\fP to read 8192 bytes from file 21,11/12.195,
1416at byte offset 24576.
1417\fIWrl\fP replies `ok'; the packet shown on the
1418second line is the first fragment of the reply, and hence is only 1472
1419bytes long (the other bytes will follow in subsequent fragments, but
1420these fragments do not have NFS or even UDP headers and so might not be
1421printed, depending on the filter expression used).
1422Because the \-v flag
1423is given, some of the file attributes (which are returned in addition
1424to the file data) are printed: the file type (``REG'', for regular file),
1425the file mode (in octal), the uid and gid, and the file size.
1426.LP
1427If the \-v flag is given more than once, even more details are printed.
1428.LP
1429Note that NFS requests are very large and much of the detail won't be printed
1430unless \fIsnaplen\fP is increased.
1431Try using `\fB\-s 192\fP' to watch
1432NFS traffic.
1433.LP
1434NFS reply packets do not explicitly identify the RPC operation.
1435Instead,
1436\fItcpdump\fP keeps track of ``recent'' requests, and matches them to the
1437replies using the transaction ID.
1438If a reply does not closely follow the
1439corresponding request, it might not be parsable.
1440.HD
1441AFS Requests and Replies
1442.LP
1443Transarc AFS (Andrew File System) requests and replies are printed
1444as:
1445.HD
1446.RS
1447.nf
1448.sp .5
1449\fIsrc.sport > dst.dport: rx packet-type\fP
1450\fIsrc.sport > dst.dport: rx packet-type service call call-name args\fP
1451\fIsrc.sport > dst.dport: rx packet-type service reply call-name args\fP
1452.sp .5
1453\f(CW
1454elvis.7001 > pike.afsfs:
1455 rx data fs call rename old fid 536876964/1/1 ".newsrc.new"
1456 new fid 536876964/1/1 ".newsrc"
1457pike.afsfs > elvis.7001: rx data fs reply rename
1458\fR
1459.sp .5
1460.fi
1461.RE
1462In the first line, host elvis sends a RX packet to pike.
1463This was
1464a RX data packet to the fs (fileserver) service, and is the start of
1465an RPC call.
1466The RPC call was a rename, with the old directory file id
1467of 536876964/1/1 and an old filename of `.newsrc.new', and a new directory
1468file id of 536876964/1/1 and a new filename of `.newsrc'.
1469The host pike
1470responds with a RPC reply to the rename call (which was successful, because
1471it was a data packet and not an abort packet).
1472.LP
1473In general, all AFS RPCs are decoded at least by RPC call name.
1474Most
1475AFS RPCs have at least some of the arguments decoded (generally only
1476the `interesting' arguments, for some definition of interesting).
1477.LP
1478The format is intended to be self-describing, but it will probably
1479not be useful to people who are not familiar with the workings of
1480AFS and RX.
1481.LP
1482If the -v (verbose) flag is given twice, acknowledgement packets and
1483additional header information is printed, such as the the RX call ID,
1484call number, sequence number, serial number, and the RX packet flags.
1485.LP
1486If the -v flag is given twice, additional information is printed,
1487such as the the RX call ID, serial number, and the RX packet flags.
1488The MTU negotiation information is also printed from RX ack packets.
1489.LP
1490If the -v flag is given three times, the security index and service id
1491are printed.
1492.LP
1493Error codes are printed for abort packets, with the exception of Ubik
1494beacon packets (because abort packets are used to signify a yes vote
1495for the Ubik protocol).
1496.LP
1497Note that AFS requests are very large and many of the arguments won't
1498be printed unless \fIsnaplen\fP is increased.
1499Try using `\fB-s 256\fP'
1500to watch AFS traffic.
1501.LP
1502AFS reply packets do not explicitly identify the RPC operation.
1503Instead,
1504\fItcpdump\fP keeps track of ``recent'' requests, and matches them to the
1505replies using the call number and service ID.
1506If a reply does not closely
1507follow the
1508corresponding request, it might not be parsable.
1509
1510.HD
1511KIP AppleTalk (DDP in UDP)
1512.LP
1513AppleTalk DDP packets encapsulated in UDP datagrams are de-encapsulated
1514and dumped as DDP packets (i.e., all the UDP header information is
1515discarded).
1516The file
1517.I /etc/atalk.names
1518is used to translate AppleTalk net and node numbers to names.
1519Lines in this file have the form
1520.RS
1521.nf
1522.sp .5
1523\fInumber name\fP
1524
1525\f(CW1.254 ether
152616.1 icsd-net
15271.254.110 ace\fR
1528.sp .5
1529.fi
1530.RE
1531The first two lines give the names of AppleTalk networks.
1532The third
1533line gives the name of a particular host (a host is distinguished
1534from a net by the 3rd octet in the number \-
1535a net number \fImust\fP have two octets and a host number \fImust\fP
1536have three octets.) The number and name should be separated by
1537whitespace (blanks or tabs).
1538The
1539.I /etc/atalk.names
1540file may contain blank lines or comment lines (lines starting with
1541a `#').
1542.LP
1543AppleTalk addresses are printed in the form
1544.RS
1545.nf
1546.sp .5
1547\fInet.host.port\fP
1548
1549\f(CW144.1.209.2 > icsd-net.112.220
1550office.2 > icsd-net.112.220
1551jssmag.149.235 > icsd-net.2\fR
1552.sp .5
1553.fi
1554.RE
1555(If the
1556.I /etc/atalk.names
1557doesn't exist or doesn't contain an entry for some AppleTalk
1558host/net number, addresses are printed in numeric form.)
1559In the first example, NBP (DDP port 2) on net 144.1 node 209
1560is sending to whatever is listening on port 220 of net icsd node 112.
1561The second line is the same except the full name of the source node
1562is known (`office').
1563The third line is a send from port 235 on
1564net jssmag node 149 to broadcast on the icsd-net NBP port (note that
1565the broadcast address (255) is indicated by a net name with no host
1566number \- for this reason it's a good idea to keep node names and
1567net names distinct in /etc/atalk.names).
1568.LP
1569NBP (name binding protocol) and ATP (AppleTalk transaction protocol)
1570packets have their contents interpreted.
1571Other protocols just dump
1572the protocol name (or number if no name is registered for the
1573protocol) and packet size.
1574
1575\fBNBP packets\fP are formatted like the following examples:
1576.RS
1577.nf
1578.sp .5
1579\s-2\f(CWicsd-net.112.220 > jssmag.2: nbp-lkup 190: "=:LaserWriter@*"
1580jssmag.209.2 > icsd-net.112.220: nbp-reply 190: "RM1140:LaserWriter@*" 250
1581techpit.2 > icsd-net.112.220: nbp-reply 190: "techpit:LaserWriter@*" 186\fR\s+2
1582.sp .5
1583.fi
1584.RE
1585The first line is a name lookup request for laserwriters sent by net icsd host
1586112 and broadcast on net jssmag.
1587The nbp id for the lookup is 190.
1588The second line shows a reply for this request (note that it has the
1589same id) from host jssmag.209 saying that it has a laserwriter
1590resource named "RM1140" registered on port 250.
1591The third line is
1592another reply to the same request saying host techpit has laserwriter
1593"techpit" registered on port 186.
1594
1595\fBATP packet\fP formatting is demonstrated by the following example:
1596.RS
1597.nf
1598.sp .5
1599\s-2\f(CWjssmag.209.165 > helios.132: atp-req 12266<0-7> 0xae030001
1600helios.132 > jssmag.209.165: atp-resp 12266:0 (512) 0xae040000
1601helios.132 > jssmag.209.165: atp-resp 12266:1 (512) 0xae040000
1602helios.132 > jssmag.209.165: atp-resp 12266:2 (512) 0xae040000
1603helios.132 > jssmag.209.165: atp-resp 12266:3 (512) 0xae040000
1604helios.132 > jssmag.209.165: atp-resp 12266:4 (512) 0xae040000
1605helios.132 > jssmag.209.165: atp-resp 12266:5 (512) 0xae040000
1606helios.132 > jssmag.209.165: atp-resp 12266:6 (512) 0xae040000
1607helios.132 > jssmag.209.165: atp-resp*12266:7 (512) 0xae040000
1608jssmag.209.165 > helios.132: atp-req 12266<3,5> 0xae030001
1609helios.132 > jssmag.209.165: atp-resp 12266:3 (512) 0xae040000
1610helios.132 > jssmag.209.165: atp-resp 12266:5 (512) 0xae040000
1611jssmag.209.165 > helios.132: atp-rel 12266<0-7> 0xae030001
1612jssmag.209.133 > helios.132: atp-req* 12267<0-7> 0xae030002\fR\s+2
1613.sp .5
1614.fi
1615.RE
1616Jssmag.209 initiates transaction id 12266 with host helios by requesting
1617up to 8 packets (the `<0-7>').
1618The hex number at the end of the line
1619is the value of the `userdata' field in the request.
1620.LP
1621Helios responds with 8 512-byte packets.
1622The `:digit' following the
1623transaction id gives the packet sequence number in the transaction
1624and the number in parens is the amount of data in the packet,
1625excluding the atp header.
1626The `*' on packet 7 indicates that the
1627EOM bit was set.
1628.LP
1629Jssmag.209 then requests that packets 3 & 5 be retransmitted.
1630Helios
1631resends them then jssmag.209 releases the transaction.
1632Finally,
1633jssmag.209 initiates the next request.
1634The `*' on the request
1635indicates that XO (`exactly once') was \fInot\fP set.
1636
1637.HD
1638IP Fragmentation
1639.LP
1640Fragmented Internet datagrams are printed as
1641.RS
1642.nf
1643.sp .5
1644\fB(frag \fIid\fB:\fIsize\fB@\fIoffset\fB+)\fR
1645\fB(frag \fIid\fB:\fIsize\fB@\fIoffset\fB)\fR
1646.sp .5
1647.fi
1648.RE
1649(The first form indicates there are more fragments.
1650The second
1651indicates this is the last fragment.)
1652.LP
1653\fIId\fP is the fragment id.
1654\fISize\fP is the fragment
1655size (in bytes) excluding the IP header.
1656\fIOffset\fP is this
1657fragment's offset (in bytes) in the original datagram.
1658.LP
1659The fragment information is output for each fragment.
1660The first
1661fragment contains the higher level protocol header and the frag
1662info is printed after the protocol info.
1663Fragments
1664after the first contain no higher level protocol header and the
1665frag info is printed after the source and destination addresses.
1666For example, here is part of an ftp from arizona.edu to lbl-rtsg.arpa
1667over a CSNET connection that doesn't appear to handle 576 byte datagrams:
1668.RS
1669.nf
1670.sp .5
1671\s-2\f(CWarizona.ftp-data > rtsg.1170: . 1024:1332(308) ack 1 win 4096 (frag 595a:328@0+)
1672arizona > rtsg: (frag 595a:204@328)
1673rtsg.1170 > arizona.ftp-data: . ack 1536 win 2560\fP\s+2
1674.sp .5
1675.fi
1676.RE
1677There are a couple of things to note here: First, addresses in the
16782nd line don't include port numbers.
1679This is because the TCP
1680protocol information is all in the first fragment and we have no idea
1681what the port or sequence numbers are when we print the later fragments.
1682Second, the tcp sequence information in the first line is printed as if there
1683were 308 bytes of user data when, in fact, there are 512 bytes (308 in
1684the first frag and 204 in the second).
1685If you are looking for holes
1686in the sequence space or trying to match up acks
1687with packets, this can fool you.
1688.LP
1689A packet with the IP \fIdon't fragment\fP flag is marked with a
1690trailing \fB(DF)\fP.
1691.HD
1692Timestamps
1693.LP
1694By default, all output lines are preceded by a timestamp.
1695The timestamp
1696is the current clock time in the form
1697.RS
1698.nf
1699\fIhh:mm:ss.frac\fP
1700.fi
1701.RE
1702and is as accurate as the kernel's clock.
1703The timestamp reflects the time the kernel first saw the packet.
1704No attempt
1705is made to account for the time lag between when the
1706Ethernet interface removed the packet from the wire and when the kernel
1707serviced the `new packet' interrupt.
1708.SH "SEE ALSO"
1709stty(1), pcap(3PCAP), bpf(4), nit(4P), pcap-savefile(@MAN_FILE_FORMATS@),
1710pcap-filter(@MAN_MISC_INFO@), pcap-tstamp-type(@MAN_MISC_INFO@)
1711.SH AUTHORS
1712The original authors are:
1713.LP
1714Van Jacobson,
1715Craig Leres and
1716Steven McCanne, all of the
1717Lawrence Berkeley National Laboratory, University of California, Berkeley, CA.
1718.LP
1719It is currently being maintained by tcpdump.org.
1720.LP
1721The current version is available via http:
1722.LP
1723.RS
1724.I http://www.tcpdump.org/
1725.RE
1726.LP
1727The original distribution is available via anonymous ftp:
1728.LP
1729.RS
1730.I ftp://ftp.ee.lbl.gov/tcpdump.tar.Z
1731.RE
1732.LP
1733IPv6/IPsec support is added by WIDE/KAME project.
1734This program uses Eric Young's SSLeay library, under specific configurations.
1735.SH BUGS
1736Please send problems, bugs, questions, desirable enhancements, patches
1737etc. to:
1738.LP
1739.RS
1740tcpdump-workers@lists.tcpdump.org
1741.RE
1742.LP
1743NIT doesn't let you watch your own outbound traffic, BPF will.
1744We recommend that you use the latter.
1745.LP
1746On Linux systems with 2.0[.x] kernels:
1747.IP
1748packets on the loopback device will be seen twice;
1749.IP
1750packet filtering cannot be done in the kernel, so that all packets must
1751be copied from the kernel in order to be filtered in user mode;
1752.IP
1753all of a packet, not just the part that's within the snapshot length,
1754will be copied from the kernel (the 2.0[.x] packet capture mechanism, if
1755asked to copy only part of a packet to userland, will not report the
1756true length of the packet; this would cause most IP packets to get an
1757error from
1758.BR tcpdump );
1759.IP
1760capturing on some PPP devices won't work correctly.
1761.LP
1762We recommend that you upgrade to a 2.2 or later kernel.
1763.LP
1764Some attempt should be made to reassemble IP fragments or, at least
1765to compute the right length for the higher level protocol.
1766.LP
1767Name server inverse queries are not dumped correctly: the (empty)
1768question section is printed rather than real query in the answer
1769section.
1770Some believe that inverse queries are themselves a bug and
1771prefer to fix the program generating them rather than \fItcpdump\fP.
1772.LP
1773A packet trace that crosses a daylight savings time change will give
1774skewed time stamps (the time change is ignored).
1775.LP
1776Filter expressions on fields other than those in Token Ring headers will
1777not correctly handle source-routed Token Ring packets.
1778.LP
1779Filter expressions on fields other than those in 802.11 headers will not
1780correctly handle 802.11 data packets with both To DS and From DS set.
1781.LP
1782.BR "ip6 proto"
1783should chase header chain, but at this moment it does not.
1784.BR "ip6 protochain"
1785is supplied for this behavior.
1786.LP
1787Arithmetic expression against transport layer headers, like \fBtcp[0]\fP,
1788does not work against IPv6 packets.
1789It only looks at IPv4 packets.
|