1 .\" @(#) $Header: /tcpdump/master/tcpdump/tcpdump.1,v 1.185.2.6 2008-05-30 01:38:21 guy Exp $ (LBL)
3 .\" $NetBSD: tcpdump.8,v 1.9 2003/03/31 00:18:17 perry Exp $
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.
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.
25 .TH TCPDUMP 1 "07 January 2008"
27 tcpdump \- dump traffic on a network
32 .B \-AdDefIKlLnNOpqRStuUvxX
94 .I spi@ipaddr algo:secret,...
104 .I postrotate-command
118 \fITcpdump\fP prints out a description of the contents of packets on a
119 network interface that match the boolean \fIexpression\fP. It can also
122 flag, which causes it to save the packet data to a file for later
123 analysis, and/or with the
125 flag, which causes it to read from a saved packet file rather than to
126 read packets from a network interface. In all cases, only packets that
133 will, if not run with the
135 flag, continue capturing packets until it is interrupted by a SIGINT
136 signal (generated, for example, by typing your interrupt character,
137 typically control-C) or a SIGTERM signal (typically generated with the
139 command); if run with the
141 flag, it will capture packets until it is interrupted by a SIGINT or
142 SIGTERM signal or the specified number of packets have been processed.
146 finishes capturing packets, it will report counts of:
148 packets ``captured'' (this is the number of packets that
150 has received and processed);
152 packets ``received by filter'' (the meaning of this depends on the OS on
155 and possibly on the way the OS was configured - if a filter was
156 specified on the command line, on some OSes it counts packets regardless
157 of whether they were matched by the filter expression and, even if they
158 were matched by the filter expression, regardless of whether
160 has read and processed them yet, on other OSes it counts only packets that were
161 matched by the filter expression regardless of whether
163 has read and processed them yet, and on other OSes it counts only
164 packets that were matched by the filter expression and were processed by
167 packets ``dropped by kernel'' (this is the number of packets that were
168 dropped, due to a lack of buffer space, by the packet capture mechanism
171 is running, if the OS reports that information to applications; if not,
172 it will be reported as 0).
174 On platforms that support the SIGINFO signal, such as most BSDs
175 (including Mac OS X) and Digital/Tru64 UNIX, it will report those counts
176 when it receives a SIGINFO signal (generated, for example, by typing
177 your ``status'' character, typically control-T, although on some
178 platforms, such as Mac OS X, the ``status'' character is not set by
179 default, so you must set it with
181 in order to use it) and will continue capturing packets.
183 Reading packets from a network interface may require that you have
184 special privileges; see the
186 man page for details. Reading a saved packet file doesn't require
191 Print each packet (minus its link level header) in ASCII. Handy for
195 Set the operating system capture buffer size to \fIbuffer_size\fP.
198 Exit after receiving \fIcount\fP packets.
201 Before writing a raw packet to a savefile, check whether the file is
202 currently larger than \fIfile_size\fP and, if so, close the current
203 savefile and open a new one. Savefiles after the first savefile will
204 have the name specified with the
206 flag, with a number after it, starting at 1 and continuing upward.
207 The units of \fIfile_size\fP are millions of bytes (1,000,000 bytes,
208 not 1,048,576 bytes).
211 Dump the compiled packet-matching code in a human readable form to
212 standard output and stop.
215 Dump packet-matching code as a
220 Dump packet-matching code as decimal numbers (preceded with a count).
223 Print the list of the network interfaces available on the system and on
226 can capture packets. For each network interface, a number and an
227 interface name, possibly followed by a text description of the
228 interface, is printed. The interface name or the number can be supplied
231 flag to specify an interface on which to capture.
233 This can be useful on systems that don't have a command to list them
234 (e.g., Windows systems, or UNIX systems lacking
235 .BR "ifconfig \-a" );
236 the number can be useful on Windows 2000 and later systems, where the
237 interface name is a somewhat complex string.
241 flag will not be supported if
243 was built with an older version of
246 .B pcap_findalldevs()
250 Print the link-level header on each dump line.
253 Use \fIspi@ipaddr algo:secret\fP for decrypting IPsec ESP packets that
254 are addressed to \fIaddr\fP and contain Security Parameter Index value
255 \fIspi\fP. This combination may be repeated with comma or newline seperation.
257 Note that setting the secret for IPv4 ESP packets is supported at this time.
264 \fBcast128-cbc\fP, or
266 The default is \fBdes-cbc\fP.
267 The ability to decrypt packets is only present if \fItcpdump\fP was compiled
268 with cryptography enabled.
270 \fIsecret\fP is the ASCII text for ESP secret key.
271 If preceeded by 0x, then a hex value will be read.
273 The option assumes RFC2406 ESP, not RFC1827 ESP.
274 The option is only for debugging purposes, and
275 the use of this option with a true `secret' key is discouraged.
276 By presenting IPsec secret key onto command line
277 you make it visible to others, via
281 In addition to the above syntax, the syntax \fIfile name\fP may be used
282 to have tcpdump read the provided file in. The file is opened upon
283 receiving the first ESP packet, so any special permissions that tcpdump
284 may have been given should already have been given up.
287 Print `foreign' IPv4 addresses numerically rather than symbolically
288 (this option is intended to get around serious brain damage in
289 Sun's NIS server \(em usually it hangs forever translating non-local
292 The test for `foreign' IPv4 addresses is done using the IPv4 address and
293 netmask of the interface on which capture is being done. If that
294 address or netmask are not available, available, either because the
295 interface on which capture is being done has no address or netmask or
296 because the capture is being done on the Linux "any" interface, which
297 can capture on more than one interface, this option will not work
301 Use \fIfile\fP as input for the filter expression.
302 An additional expression given on the command line is ignored.
305 If specified, rotates the dump file specified with the
307 option every \fIrotate_seconds\fP seconds.
308 Savefiles will have the name specified by
310 which should include a time format as defined by
312 If no time format is specified, each new file will overwrite the previous.
314 If used in conjunction with the
316 option, filenames will take the form of `\fIfile\fP<count>'.
319 Listen on \fIinterface\fP.
320 If unspecified, \fItcpdump\fP searches the system interface list for the
321 lowest numbered, configured up interface (excluding loopback).
322 Ties are broken by choosing the earliest match.
324 On Linux systems with 2.2 or later kernels, an
326 argument of ``any'' can be used to capture packets from all interfaces.
327 Note that captures on the ``any'' device will not be done in promiscuous
332 flag is supported, an interface number as printed by that flag can be
338 Put the interface in "monitor mode"; this is supported only on IEEE
339 802.11 Wi-Fi interfaces, and supported only on some operating systems.
341 Note that in monitor mode the adapter might disassociate from the
342 network with which it's associated, so that you will not be able to use
343 any wireless networks with that adapter. This could prevent accessing
344 files on a network server, or resolving host names or network addresses,
345 if you are capturing in monitor mode and are not connected to another
346 network with another adapter.
349 Don't attempt to verify TCP checksums. This is useful for interfaces
350 that perform the TCP checksum calculation in hardware; otherwise,
351 all outgoing TCP checksums will be flagged as bad.
354 Make stdout line buffered.
355 Useful if you want to see the data
359 ``tcpdump\ \ \-l\ \ |\ \ tee dat'' or
360 ``tcpdump\ \ \-l \ \ > dat\ \ &\ \ tail\ \ \-f\ \ dat''.
363 List the known data link types for the interface and exit.
366 Load SMI MIB module definitions from file \fImodule\fR.
368 can be used several times to load several MIB modules into \fItcpdump\fP.
371 Use \fIsecret\fP as a shared secret for validating the digests found in
372 TCP segments with the TCP-MD5 option (RFC 2385), if present.
375 Don't convert addresses (i.e., host addresses, port numbers, etc.) to names.
378 Don't print domain name qualification of host names.
380 if you give this flag then \fItcpdump\fP will print ``nic''
381 instead of ``nic.ddn.mil''.
384 Do not run the packet-matching code optimizer.
386 if you suspect a bug in the optimizer.
389 \fIDon't\fP put the interface
390 into promiscuous mode.
391 Note that the interface might be in promiscuous
392 mode for some other reason; hence, `-p' cannot be used as an abbreviation for
393 `ether host {local-hw-addr} or ether broadcast'.
396 Quick (quiet?) output.
397 Print less protocol information so output
401 Assume ESP/AH packets to be based on old specification (RFC1825 to RFC1829).
402 If specified, \fItcpdump\fP will not print replay prevention field.
403 Since there is no protocol version field in ESP/AH specification,
404 \fItcpdump\fP cannot deduce the version of ESP/AH protocol.
407 Read packets from \fIfile\fR (which was created with the
410 Standard input is used if \fIfile\fR is ``-''.
413 Print absolute, rather than relative, TCP sequence numbers.
416 Snarf \fIsnaplen\fP bytes of data from each packet rather than the
417 default of 68 (with SunOS's NIT, the minimum is actually 96).
418 68 bytes is adequate for IP, ICMP, TCP
419 and UDP but may truncate protocol information from name server and NFS
421 Packets truncated because of a limited snapshot
422 are indicated in the output with ``[|\fIproto\fP]'', where \fIproto\fP
423 is the name of the protocol level at which the truncation has occurred.
424 Note that taking larger snapshots both increases
425 the amount of time it takes to process packets and, effectively,
426 decreases the amount of packet buffering.
427 This may cause packets to be
429 You should limit \fIsnaplen\fP to the smallest number that will
430 capture the protocol information you're interested in.
432 \fIsnaplen\fP to 0 means use the required length to catch whole packets.
435 Force packets selected by "\fIexpression\fP" to be interpreted the
436 specified \fItype\fR.
437 Currently known types are
438 \fBaodv\fR (Ad-hoc On-demand Distance Vector protocol),
439 \fBcnfp\fR (Cisco NetFlow protocol),
440 \fBrpc\fR (Remote Procedure Call),
441 \fBrtp\fR (Real-Time Applications protocol),
442 \fBrtcp\fR (Real-Time Applications control protocol),
443 \fBsnmp\fR (Simple Network Management Protocol),
444 \fBtftp\fR (Trivial File Transfer Protocol),
445 \fBvat\fR (Visual Audio Tool),
447 \fBwb\fR (distributed White Board).
450 \fIDon't\fP print a timestamp on each dump line.
453 Print an unformatted timestamp on each dump line.
456 Print a delta (micro-second resolution) between current and previous line
460 Print a timestamp in default format proceeded by date on each dump line.
463 Print a delta (micro-second resolution) between current and first line
467 Print undecoded NFS handles.
470 Make output saved via the
472 option ``packet-buffered''; i.e., as each packet is saved, it will be
473 written to the output file, rather than being written only when the
478 flag will not be supported if
480 was built with an older version of
487 When parsing and printing, produce (slightly more) verbose output.
488 For example, the time to live,
489 identification, total length and options in an IP packet are printed.
490 Also enables additional packet integrity checks such as verifying the
491 IP and ICMP header checksum.
493 When writing to a file with the
495 option, report, every 10 seconds, the number of packets captured.
498 Even more verbose output.
499 For example, additional fields are
500 printed from NFS reply packets, and SMB packets are fully decoded.
503 Even more verbose output.
505 telnet \fBSB\fP ... \fBSE\fP options
509 Telnet options are printed in hex as well.
512 Write the raw packets to \fIfile\fR rather than parsing and printing
514 They can later be printed with the \-r option.
515 Standard output is used if \fIfile\fR is ``-''.
518 Used in conjunction with the
520 option, this will limit the number
521 of files created to the specified number, and begin overwriting files
522 from the beginning, thus creating a 'rotating' buffer.
523 In addition, it will name
524 the files with enough leading 0s to support the maximum number of
525 files, allowing them to sort correctly.
527 Used in conjunction with the
529 option, this will limit the number of rotated dump files that get
530 created, exiting with status 0 when reaching the limit. If used with
532 as well, the behavior will result in cyclical files per timeslice.
535 When parsing and printing,
536 in addition to printing the headers of each packet, print the data of
537 each packet (minus its link level header) in hex.
538 The smaller of the entire packet or
540 bytes will be printed. Note that this is the entire link-layer
541 packet, so for link layers that pad (e.g. Ethernet), the padding bytes
542 will also be printed when the higher layer packet is shorter than the
546 When parsing and printing,
547 in addition to printing the headers of each packet, print the data of
550 its link level header, in hex.
553 When parsing and printing,
554 in addition to printing the headers of each packet, print the data of
555 each packet (minus its link level header) in hex and ASCII.
556 This is very handy for analysing new protocols.
559 When parsing and printing,
560 in addition to printing the headers of each packet, print the data of
563 its link level header, in hex and ASCII.
566 Set the data link type to use while capturing packets to \fIdatalinktype\fP.
569 Used in conjunction with the
573 options, this will make
579 is the savefile being closed after each rotation. For example, specifying
583 will compress each savefile using gzip or bzip2.
585 Note that tcpdump will run the command in parallel to the capture, using
586 the lowest priority so that this doesn't disturb the capture process.
588 And in case you would like to use a command that itself takes flags or
589 different arguments, you can always write a shell script that will take the
590 savefile name as the only argument, make the flags & arguments arrangements
591 and execute the command that you want.
594 Drops privileges (if root) and changes user ID to
596 and the group ID to the primary group of
599 This behavior can also be enabled by default at compile time.
600 .IP "\fI expression\fP"
602 selects which packets will be dumped.
603 If no \fIexpression\fP
604 is given, all packets on the net will be dumped.
606 only packets for which \fIexpression\fP is `true' will be dumped.
608 For the \fIexpression\fP syntax, see
611 Expression arguments can be passed to \fItcpdump\fP as either a single
612 argument or as multiple arguments, whichever is more convenient.
613 Generally, if the expression contains Shell metacharacters, it is
614 easier to pass it as a single, quoted argument.
615 Multiple arguments are concatenated with spaces before being parsed.
618 To print all packets arriving at or departing from \fIsundown\fP:
621 \fBtcpdump host sundown\fP
625 To print traffic between \fIhelios\fR and either \fIhot\fR or \fIace\fR:
628 \fBtcpdump host helios and \\( hot or ace \\)\fP
632 To print all IP packets between \fIace\fR and any host except \fIhelios\fR:
635 \fBtcpdump ip host ace and not helios\fP
639 To print all traffic between local hosts and hosts at Berkeley:
643 tcpdump net ucb-ether
647 To print all ftp traffic through internet gateway \fIsnup\fP:
648 (note that the expression is quoted to prevent the shell from
649 (mis-)interpreting the parentheses):
653 tcpdump 'gateway snup and (port ftp or ftp-data)'
657 To print traffic neither sourced from nor destined for local hosts
658 (if you gateway to one other net, this stuff should never make it
659 onto your local net).
663 tcpdump ip and not net \fIlocalnet\fP
667 To print the start and end packets (the SYN and FIN packets) of each
668 TCP conversation that involves a non-local host.
672 tcpdump 'tcp[tcpflags] & (tcp-syn|tcp-fin) != 0 and not src and dst net \fIlocalnet\fP'
676 To print all IPv4 HTTP packets to and from port 80, i.e. print only
677 packets that contain data, not, for example, SYN and FIN packets and
678 ACK-only packets. (IPv6 is left as an exercise for the reader.)
682 tcpdump 'tcp port 80 and (((ip[2:2] - ((ip[0]&0xf)<<2)) - ((tcp[12]&0xf0)>>2)) != 0)'
686 To print IP packets longer than 576 bytes sent through gateway \fIsnup\fP:
690 tcpdump 'gateway snup and ip[2:2] > 576'
694 To print IP broadcast or multicast packets that were
696 sent via Ethernet broadcast or multicast:
700 tcpdump 'ether[0] & 1 = 0 and ip[16] >= 224'
704 To print all ICMP packets that are not echo requests/replies (i.e., not
709 tcpdump 'icmp[icmptype] != icmp-echo and icmp[icmptype] != icmp-echoreply'
714 The output of \fItcpdump\fP is protocol dependent.
716 gives a brief description and examples of most of the formats.
724 If the '-e' option is given, the link level header is printed out.
725 On Ethernets, the source and destination addresses, protocol,
726 and packet length are printed.
728 On FDDI networks, the '-e' option causes \fItcpdump\fP to print
729 the `frame control' field, the source and destination addresses,
730 and the packet length.
731 (The `frame control' field governs the
732 interpretation of the rest of the packet.
734 as those containing IP datagrams) are `async' packets, with a priority
735 value between 0 and 7; for example, `\fBasync4\fR'.
737 are assumed to contain an 802.2 Logical Link Control (LLC) packet;
738 the LLC header is printed if it is \fInot\fR an ISO datagram or a
739 so-called SNAP packet.
741 On Token Ring networks, the '-e' option causes \fItcpdump\fP to print
742 the `access control' and `frame control' fields, the source and
743 destination addresses, and the packet length.
745 packets are assumed to contain an LLC packet.
746 Regardless of whether
747 the '-e' option is specified or not, the source routing information is
748 printed for source-routed packets.
750 On 802.11 networks, the '-e' option causes \fItcpdump\fP to print
751 the `frame control' fields, all of the addresses in the 802.11 header,
752 and the packet length.
754 packets are assumed to contain an LLC packet.
756 \fI(N.B.: The following description assumes familiarity with
757 the SLIP compression algorithm described in RFC-1144.)\fP
759 On SLIP links, a direction indicator (``I'' for inbound, ``O'' for outbound),
760 packet type, and compression information are printed out.
761 The packet type is printed first.
762 The three types are \fIip\fP, \fIutcp\fP, and \fIctcp\fP.
763 No further link information is printed for \fIip\fR packets.
764 For TCP packets, the connection identifier is printed following the type.
765 If the packet is compressed, its encoded header is printed out.
766 The special cases are printed out as
767 \fB*S+\fIn\fR and \fB*SA+\fIn\fR, where \fIn\fR is the amount by which
768 the sequence number (or sequence number and ack) has changed.
769 If it is not a special case,
770 zero or more changes are printed.
771 A change is indicated by U (urgent pointer), W (window), A (ack),
772 S (sequence number), and I (packet ID), followed by a delta (+n or -n),
774 Finally, the amount of data in the packet and compressed header length
777 For example, the following line shows an outbound compressed TCP packet,
778 with an implicit connection identifier; the ack has changed by 6,
779 the sequence number by 49, and the packet ID by 6; there are 3 bytes of
780 data and 6 bytes of compressed header:
783 \fBO ctcp * A+6 S+49 I+6 3 (6)\fP
789 Arp/rarp output shows the type of request and its arguments.
791 format is intended to be self explanatory.
792 Here is a short sample taken from the start of an `rlogin' from
793 host \fIrtsg\fP to host \fIcsam\fP:
797 \f(CWarp who-has csam tell rtsg
798 arp reply csam is-at CSAM\fR
802 The first line says that rtsg sent an arp packet asking
803 for the Ethernet address of internet host csam.
805 replies with its Ethernet address (in this example, Ethernet addresses
806 are in caps and internet addresses in lower case).
808 This would look less redundant if we had done \fItcpdump \-n\fP:
812 \f(CWarp who-has 128.3.254.6 tell 128.3.254.68
813 arp reply 128.3.254.6 is-at 02:07:01:00:01:c4\fP
817 If we had done \fItcpdump \-e\fP, the fact that the first packet is
818 broadcast and the second is point-to-point would be visible:
822 \f(CWRTSG Broadcast 0806 64: arp who-has csam tell rtsg
823 CSAM RTSG 0806 64: arp reply csam is-at CSAM\fR
827 For the first packet this says the Ethernet source address is RTSG, the
828 destination is the Ethernet broadcast address, the type field
829 contained hex 0806 (type ETHER_ARP) and the total length was 64 bytes.
833 \fI(N.B.:The following description assumes familiarity with
834 the TCP protocol described in RFC-793.
835 If you are not familiar
836 with the protocol, neither this description nor \fItcpdump\fP will
837 be of much use to you.)\fP
839 The general format of a tcp protocol line is:
843 \fIsrc > dst: flags data-seqno ack window urgent options\fP
847 \fISrc\fP and \fIdst\fP are the source and destination IP
849 \fIFlags\fP are some combination of S (SYN),
850 F (FIN), P (PUSH), R (RST), W (ECN CWR) or E (ECN-Echo), or a single
852 \fIData-seqno\fP describes the portion of sequence space covered
853 by the data in this packet (see example below).
854 \fIAck\fP is sequence number of the next data expected the other
855 direction on this connection.
856 \fIWindow\fP is the number of bytes of receive buffer space available
857 the other direction on this connection.
858 \fIUrg\fP indicates there is `urgent' data in the packet.
859 \fIOptions\fP are tcp options enclosed in angle brackets (e.g., <mss 1024>).
861 \fISrc, dst\fP and \fIflags\fP are always present.
863 depend on the contents of the packet's tcp protocol header and
864 are output only if appropriate.
866 Here is the opening portion of an rlogin from host \fIrtsg\fP to
871 \s-2\f(CWrtsg.1023 > csam.login: S 768512:768512(0) win 4096 <mss 1024>
872 csam.login > rtsg.1023: S 947648:947648(0) ack 768513 win 4096 <mss 1024>
873 rtsg.1023 > csam.login: . ack 1 win 4096
874 rtsg.1023 > csam.login: P 1:2(1) ack 1 win 4096
875 csam.login > rtsg.1023: . ack 2 win 4096
876 rtsg.1023 > csam.login: P 2:21(19) ack 1 win 4096
877 csam.login > rtsg.1023: P 1:2(1) ack 21 win 4077
878 csam.login > rtsg.1023: P 2:3(1) ack 21 win 4077 urg 1
879 csam.login > rtsg.1023: P 3:4(1) ack 21 win 4077 urg 1\fR\s+2
883 The first line says that tcp port 1023 on rtsg sent a packet
886 The \fBS\fP indicates that the \fISYN\fP flag was set.
887 The packet sequence number was 768512 and it contained no data.
888 (The notation is `first:last(nbytes)' which means `sequence
890 up to but not including \fIlast\fP which is \fInbytes\fP bytes of user data'.)
891 There was no piggy-backed ack, the available receive window was 4096
892 bytes and there was a max-segment-size option requesting an mss of
895 Csam replies with a similar packet except it includes a piggy-backed
897 Rtsg then acks csam's SYN.
900 The packet contained no data so there is no data sequence number.
901 Note that the ack sequence
902 number is a small integer (1).
903 The first time \fItcpdump\fP sees a
904 tcp `conversation', it prints the sequence number from the packet.
905 On subsequent packets of the conversation, the difference between
906 the current packet's sequence number and this initial sequence number
908 This means that sequence numbers after the
909 first can be interpreted
910 as relative byte positions in the conversation's data stream (with the
911 first data byte each direction being `1').
912 `-S' will override this
913 feature, causing the original sequence numbers to be output.
915 On the 6th line, rtsg sends csam 19 bytes of data (bytes 2 through 20
916 in the rtsg \(-> csam side of the conversation).
917 The PUSH flag is set in the packet.
918 On the 7th line, csam says it's received data sent by rtsg up to
919 but not including byte 21.
920 Most of this data is apparently sitting in the
921 socket buffer since csam's receive window has gotten 19 bytes smaller.
922 Csam also sends one byte of data to rtsg in this packet.
923 On the 8th and 9th lines,
924 csam sends two bytes of urgent, pushed data to rtsg.
926 If the snapshot was small enough that \fItcpdump\fP didn't capture
927 the full TCP header, it interprets as much of the header as it can
928 and then reports ``[|\fItcp\fP]'' to indicate the remainder could not
930 If the header contains a bogus option (one with a length
931 that's either too small or beyond the end of the header), \fItcpdump\fP
932 reports it as ``[\fIbad opt\fP]'' and does not interpret any further
933 options (since it's impossible to tell where they start).
935 length indicates options are present but the IP datagram length is not
936 long enough for the options to actually be there, \fItcpdump\fP reports
937 it as ``[\fIbad hdr length\fP]''.
939 .B Capturing TCP packets with particular flag combinations (SYN-ACK, URG-ACK, etc.)
941 There are 8 bits in the control bits section of the TCP header:
943 .I CWR | ECE | URG | ACK | PSH | RST | SYN | FIN
945 Let's assume that we want to watch packets used in establishing
947 Recall that TCP uses a 3-way handshake protocol
948 when it initializes a new connection; the connection sequence with
949 regard to the TCP control bits is
955 2) Recipient responds with SYN, ACK
961 Now we're interested in capturing packets that have only the
962 SYN bit set (Step 1).
963 Note that we don't want packets from step 2
964 (SYN-ACK), just a plain initial SYN.
965 What we need is a correct filter
966 expression for \fItcpdump\fP.
968 Recall the structure of a TCP header without options:
972 -----------------------------------------------------------------
973 | source port | destination port |
974 -----------------------------------------------------------------
976 -----------------------------------------------------------------
977 | acknowledgment number |
978 -----------------------------------------------------------------
979 | HL | rsvd |C|E|U|A|P|R|S|F| window size |
980 -----------------------------------------------------------------
981 | TCP checksum | urgent pointer |
982 -----------------------------------------------------------------
985 A TCP header usually holds 20 octets of data, unless options are
987 The first line of the graph contains octets 0 - 3, the
988 second line shows octets 4 - 7 etc.
990 Starting to count with 0, the relevant TCP control bits are contained
995 ----------------|---------------|---------------|----------------
996 | HL | rsvd |C|E|U|A|P|R|S|F| window size |
997 ----------------|---------------|---------------|----------------
1001 Let's have a closer look at octet no. 13:
1011 These are the TCP control bits we are interested
1013 We have numbered the bits in this octet from 0 to 7, right to
1014 left, so the PSH bit is bit number 3, while the URG bit is number 5.
1016 Recall that we want to capture packets with only SYN set.
1017 Let's see what happens to octet 13 if a TCP datagram arrives
1018 with the SYN bit set in its header:
1029 control bits section we see that only bit number 1 (SYN) is set.
1031 Assuming that octet number 13 is an 8-bit unsigned integer in
1032 network byte order, the binary value of this octet is
1036 and its decimal representation is
1040 0*2 + 0*2 + 0*2 + 0*2 + 0*2 + 0*2 + 1*2 + 0*2 = 2
1043 We're almost done, because now we know that if only SYN is set,
1044 the value of the 13th octet in the TCP header, when interpreted
1045 as a 8-bit unsigned integer in network byte order, must be exactly 2.
1047 This relationship can be expressed as
1053 We can use this expression as the filter for \fItcpdump\fP in order
1054 to watch packets which have only SYN set:
1057 tcpdump -i xl0 tcp[13] == 2
1060 The expression says "let the 13th octet of a TCP datagram have
1061 the decimal value 2", which is exactly what we want.
1063 Now, let's assume that we need to capture SYN packets, but we
1064 don't care if ACK or any other TCP control bit is set at the
1066 Let's see what happens to octet 13 when a TCP datagram
1067 with SYN-ACK set arrives:
1077 Now bits 1 and 4 are set in the 13th octet.
1083 which translates to decimal
1087 0*2 + 0*2 + 0*2 + 1*2 + 0*2 + 0*2 + 1*2 + 0*2 = 18
1090 Now we can't just use 'tcp[13] == 18' in the \fItcpdump\fP filter
1091 expression, because that would select only those packets that have
1092 SYN-ACK set, but not those with only SYN set.
1093 Remember that we don't care
1094 if ACK or any other control bit is set as long as SYN is set.
1096 In order to achieve our goal, we need to logically AND the
1097 binary value of octet 13 with some other value to preserve
1099 We know that we want SYN to be set in any case,
1100 so we'll logically AND the value in the 13th octet with
1101 the binary value of a SYN:
1105 00010010 SYN-ACK 00000010 SYN
1106 AND 00000010 (we want SYN) AND 00000010 (we want SYN)
1108 = 00000010 = 00000010
1111 We see that this AND operation delivers the same result
1112 regardless whether ACK or another TCP control bit is set.
1113 The decimal representation of the AND value as well as
1114 the result of this operation is 2 (binary 00000010),
1115 so we know that for packets with SYN set the following
1116 relation must hold true:
1118 ( ( value of octet 13 ) AND ( 2 ) ) == ( 2 )
1120 This points us to the \fItcpdump\fP filter expression
1123 tcpdump -i xl0 'tcp[13] & 2 == 2'
1126 Note that you should use single quotes or a backslash
1127 in the expression to hide the AND ('&') special character
1133 UDP format is illustrated by this rwho packet:
1137 \f(CWactinide.who > broadcast.who: udp 84\fP
1141 This says that port \fIwho\fP on host \fIactinide\fP sent a udp
1142 datagram to port \fIwho\fP on host \fIbroadcast\fP, the Internet
1144 The packet contained 84 bytes of user data.
1146 Some UDP services are recognized (from the source or destination
1147 port number) and the higher level protocol information printed.
1148 In particular, Domain Name service requests (RFC-1034/1035) and Sun
1149 RPC calls (RFC-1050) to NFS.
1151 UDP Name Server Requests
1153 \fI(N.B.:The following description assumes familiarity with
1154 the Domain Service protocol described in RFC-1035.
1155 If you are not familiar
1156 with the protocol, the following description will appear to be written
1159 Name server requests are formatted as
1163 \fIsrc > dst: id op? flags qtype qclass name (len)\fP
1165 \f(CWh2opolo.1538 > helios.domain: 3+ A? ucbvax.berkeley.edu. (37)\fR
1169 Host \fIh2opolo\fP asked the domain server on \fIhelios\fP for an
1170 address record (qtype=A) associated with the name \fIucbvax.berkeley.edu.\fP
1171 The query id was `3'.
1172 The `+' indicates the \fIrecursion desired\fP flag
1174 The query length was 37 bytes, not including the UDP and
1175 IP protocol headers.
1176 The query operation was the normal one, \fIQuery\fP,
1177 so the op field was omitted.
1178 If the op had been anything else, it would
1179 have been printed between the `3' and the `+'.
1180 Similarly, the qclass was the normal one,
1181 \fIC_IN\fP, and omitted.
1182 Any other qclass would have been printed
1183 immediately after the `A'.
1185 A few anomalies are checked and may result in extra fields enclosed in
1186 square brackets: If a query contains an answer, authority records or
1187 additional records section,
1192 are printed as `[\fIn\fPa]', `[\fIn\fPn]' or `[\fIn\fPau]' where \fIn\fP
1193 is the appropriate count.
1194 If any of the response bits are set (AA, RA or rcode) or any of the
1195 `must be zero' bits are set in bytes two and three, `[b2&3=\fIx\fP]'
1196 is printed, where \fIx\fP is the hex value of header bytes two and three.
1198 UDP Name Server Responses
1200 Name server responses are formatted as
1204 \fIsrc > dst: id op rcode flags a/n/au type class data (len)\fP
1206 \f(CWhelios.domain > h2opolo.1538: 3 3/3/7 A 128.32.137.3 (273)
1207 helios.domain > h2opolo.1537: 2 NXDomain* 0/1/0 (97)\fR
1211 In the first example, \fIhelios\fP responds to query id 3 from \fIh2opolo\fP
1212 with 3 answer records, 3 name server records and 7 additional records.
1213 The first answer record is type A (address) and its data is internet
1214 address 128.32.137.3.
1215 The total size of the response was 273 bytes,
1216 excluding UDP and IP headers.
1217 The op (Query) and response code
1218 (NoError) were omitted, as was the class (C_IN) of the A record.
1220 In the second example, \fIhelios\fP responds to query 2 with a
1221 response code of non-existent domain (NXDomain) with no answers,
1222 one name server and no authority records.
1223 The `*' indicates that
1224 the \fIauthoritative answer\fP bit was set.
1226 answers, no type, class or data were printed.
1228 Other flag characters that might appear are `\-' (recursion available,
1229 RA, \fInot\fP set) and `|' (truncated message, TC, set).
1231 `question' section doesn't contain exactly one entry, `[\fIn\fPq]'
1234 Note that name server requests and responses tend to be large and the
1235 default \fIsnaplen\fP of 68 bytes may not capture enough of the packet
1237 Use the \fB\-s\fP flag to increase the snaplen if you
1238 need to seriously investigate name server traffic.
1240 has worked well for me.
1245 \fItcpdump\fP now includes fairly extensive SMB/CIFS/NBT decoding for data
1246 on UDP/137, UDP/138 and TCP/139.
1247 Some primitive decoding of IPX and
1248 NetBEUI SMB data is also done.
1250 By default a fairly minimal decode is done, with a much more detailed
1251 decode done if -v is used.
1252 Be warned that with -v a single SMB packet
1253 may take up a page or more, so only use -v if you really want all the
1256 For information on SMB packet formats and what all te fields mean see
1257 www.cifs.org or the pub/samba/specs/ directory on your favorite
1258 samba.org mirror site.
1259 The SMB patches were written by Andrew Tridgell
1263 NFS Requests and Replies
1265 Sun NFS (Network File System) requests and replies are printed as:
1269 \fIsrc.xid > dst.nfs: len op args\fP
1270 \fIsrc.nfs > dst.xid: reply stat len op results\fP
1273 sushi.6709 > wrl.nfs: 112 readlink fh 21,24/10.73165
1274 wrl.nfs > sushi.6709: reply ok 40 readlink "../var"
1275 sushi.201b > wrl.nfs:
1276 144 lookup fh 9,74/4096.6878 "xcolors"
1277 wrl.nfs > sushi.201b:
1278 reply ok 128 lookup fh 9,74/4134.3150
1283 In the first line, host \fIsushi\fP sends a transaction with id \fI6709\fP
1284 to \fIwrl\fP (note that the number following the src host is a
1285 transaction id, \fInot\fP the source port).
1286 The request was 112 bytes,
1287 excluding the UDP and IP headers.
1288 The operation was a \fIreadlink\fP
1289 (read symbolic link) on file handle (\fIfh\fP) 21,24/10.731657119.
1290 (If one is lucky, as in this case, the file handle can be interpreted
1291 as a major,minor device number pair, followed by the inode number and
1293 \fIWrl\fP replies `ok' with the contents of the link.
1295 In the third line, \fIsushi\fP asks \fIwrl\fP to lookup the name
1296 `\fIxcolors\fP' in directory file 9,74/4096.6878.
1297 Note that the data printed
1298 depends on the operation type.
1299 The format is intended to be self
1300 explanatory if read in conjunction with
1301 an NFS protocol spec.
1303 If the \-v (verbose) flag is given, additional information is printed.
1309 sushi.1372a > wrl.nfs:
1310 148 read fh 21,11/12.195 8192 bytes @ 24576
1311 wrl.nfs > sushi.1372a:
1312 reply ok 1472 read REG 100664 ids 417/0 sz 29388
1317 (\-v also prints the IP header TTL, ID, length, and fragmentation fields,
1318 which have been omitted from this example.) In the first line,
1319 \fIsushi\fP asks \fIwrl\fP to read 8192 bytes from file 21,11/12.195,
1320 at byte offset 24576.
1321 \fIWrl\fP replies `ok'; the packet shown on the
1322 second line is the first fragment of the reply, and hence is only 1472
1323 bytes long (the other bytes will follow in subsequent fragments, but
1324 these fragments do not have NFS or even UDP headers and so might not be
1325 printed, depending on the filter expression used).
1326 Because the \-v flag
1327 is given, some of the file attributes (which are returned in addition
1328 to the file data) are printed: the file type (``REG'', for regular file),
1329 the file mode (in octal), the uid and gid, and the file size.
1331 If the \-v flag is given more than once, even more details are printed.
1333 Note that NFS requests are very large and much of the detail won't be printed
1334 unless \fIsnaplen\fP is increased.
1335 Try using `\fB\-s 192\fP' to watch
1338 NFS reply packets do not explicitly identify the RPC operation.
1340 \fItcpdump\fP keeps track of ``recent'' requests, and matches them to the
1341 replies using the transaction ID.
1342 If a reply does not closely follow the
1343 corresponding request, it might not be parsable.
1345 AFS Requests and Replies
1347 Transarc AFS (Andrew File System) requests and replies are printed
1353 \fIsrc.sport > dst.dport: rx packet-type\fP
1354 \fIsrc.sport > dst.dport: rx packet-type service call call-name args\fP
1355 \fIsrc.sport > dst.dport: rx packet-type service reply call-name args\fP
1358 elvis.7001 > pike.afsfs:
1359 rx data fs call rename old fid 536876964/1/1 ".newsrc.new"
1360 new fid 536876964/1/1 ".newsrc"
1361 pike.afsfs > elvis.7001: rx data fs reply rename
1366 In the first line, host elvis sends a RX packet to pike.
1368 a RX data packet to the fs (fileserver) service, and is the start of
1370 The RPC call was a rename, with the old directory file id
1371 of 536876964/1/1 and an old filename of `.newsrc.new', and a new directory
1372 file id of 536876964/1/1 and a new filename of `.newsrc'.
1374 responds with a RPC reply to the rename call (which was successful, because
1375 it was a data packet and not an abort packet).
1377 In general, all AFS RPCs are decoded at least by RPC call name.
1379 AFS RPCs have at least some of the arguments decoded (generally only
1380 the `interesting' arguments, for some definition of interesting).
1382 The format is intended to be self-describing, but it will probably
1383 not be useful to people who are not familiar with the workings of
1386 If the -v (verbose) flag is given twice, acknowledgement packets and
1387 additional header information is printed, such as the the RX call ID,
1388 call number, sequence number, serial number, and the RX packet flags.
1390 If the -v flag is given twice, additional information is printed,
1391 such as the the RX call ID, serial number, and the RX packet flags.
1392 The MTU negotiation information is also printed from RX ack packets.
1394 If the -v flag is given three times, the security index and service id
1397 Error codes are printed for abort packets, with the exception of Ubik
1398 beacon packets (because abort packets are used to signify a yes vote
1399 for the Ubik protocol).
1401 Note that AFS requests are very large and many of the arguments won't
1402 be printed unless \fIsnaplen\fP is increased.
1403 Try using `\fB-s 256\fP'
1404 to watch AFS traffic.
1406 AFS reply packets do not explicitly identify the RPC operation.
1408 \fItcpdump\fP keeps track of ``recent'' requests, and matches them to the
1409 replies using the call number and service ID.
1410 If a reply does not closely
1412 corresponding request, it might not be parsable.
1415 KIP AppleTalk (DDP in UDP)
1417 AppleTalk DDP packets encapsulated in UDP datagrams are de-encapsulated
1418 and dumped as DDP packets (i.e., all the UDP header information is
1422 is used to translate AppleTalk net and node numbers to names.
1423 Lines in this file have the form
1435 The first two lines give the names of AppleTalk networks.
1437 line gives the name of a particular host (a host is distinguished
1438 from a net by the 3rd octet in the number \-
1439 a net number \fImust\fP have two octets and a host number \fImust\fP
1440 have three octets.) The number and name should be separated by
1441 whitespace (blanks or tabs).
1444 file may contain blank lines or comment lines (lines starting with
1447 AppleTalk addresses are printed in the form
1453 \f(CW144.1.209.2 > icsd-net.112.220
1454 office.2 > icsd-net.112.220
1455 jssmag.149.235 > icsd-net.2\fR
1461 doesn't exist or doesn't contain an entry for some AppleTalk
1462 host/net number, addresses are printed in numeric form.)
1463 In the first example, NBP (DDP port 2) on net 144.1 node 209
1464 is sending to whatever is listening on port 220 of net icsd node 112.
1465 The second line is the same except the full name of the source node
1466 is known (`office').
1467 The third line is a send from port 235 on
1468 net jssmag node 149 to broadcast on the icsd-net NBP port (note that
1469 the broadcast address (255) is indicated by a net name with no host
1470 number \- for this reason it's a good idea to keep node names and
1471 net names distinct in /etc/atalk.names).
1473 NBP (name binding protocol) and ATP (AppleTalk transaction protocol)
1474 packets have their contents interpreted.
1475 Other protocols just dump
1476 the protocol name (or number if no name is registered for the
1477 protocol) and packet size.
1479 \fBNBP packets\fP are formatted like the following examples:
1483 \s-2\f(CWicsd-net.112.220 > jssmag.2: nbp-lkup 190: "=:LaserWriter@*"
1484 jssmag.209.2 > icsd-net.112.220: nbp-reply 190: "RM1140:LaserWriter@*" 250
1485 techpit.2 > icsd-net.112.220: nbp-reply 190: "techpit:LaserWriter@*" 186\fR\s+2
1489 The first line is a name lookup request for laserwriters sent by net icsd host
1490 112 and broadcast on net jssmag.
1491 The nbp id for the lookup is 190.
1492 The second line shows a reply for this request (note that it has the
1493 same id) from host jssmag.209 saying that it has a laserwriter
1494 resource named "RM1140" registered on port 250.
1496 another reply to the same request saying host techpit has laserwriter
1497 "techpit" registered on port 186.
1499 \fBATP packet\fP formatting is demonstrated by the following example:
1503 \s-2\f(CWjssmag.209.165 > helios.132: atp-req 12266<0-7> 0xae030001
1504 helios.132 > jssmag.209.165: atp-resp 12266:0 (512) 0xae040000
1505 helios.132 > jssmag.209.165: atp-resp 12266:1 (512) 0xae040000
1506 helios.132 > jssmag.209.165: atp-resp 12266:2 (512) 0xae040000
1507 helios.132 > jssmag.209.165: atp-resp 12266:3 (512) 0xae040000
1508 helios.132 > jssmag.209.165: atp-resp 12266:4 (512) 0xae040000
1509 helios.132 > jssmag.209.165: atp-resp 12266:5 (512) 0xae040000
1510 helios.132 > jssmag.209.165: atp-resp 12266:6 (512) 0xae040000
1511 helios.132 > jssmag.209.165: atp-resp*12266:7 (512) 0xae040000
1512 jssmag.209.165 > helios.132: atp-req 12266<3,5> 0xae030001
1513 helios.132 > jssmag.209.165: atp-resp 12266:3 (512) 0xae040000
1514 helios.132 > jssmag.209.165: atp-resp 12266:5 (512) 0xae040000
1515 jssmag.209.165 > helios.132: atp-rel 12266<0-7> 0xae030001
1516 jssmag.209.133 > helios.132: atp-req* 12267<0-7> 0xae030002\fR\s+2
1520 Jssmag.209 initiates transaction id 12266 with host helios by requesting
1521 up to 8 packets (the `<0-7>').
1522 The hex number at the end of the line
1523 is the value of the `userdata' field in the request.
1525 Helios responds with 8 512-byte packets.
1526 The `:digit' following the
1527 transaction id gives the packet sequence number in the transaction
1528 and the number in parens is the amount of data in the packet,
1529 excluding the atp header.
1530 The `*' on packet 7 indicates that the
1533 Jssmag.209 then requests that packets 3 & 5 be retransmitted.
1535 resends them then jssmag.209 releases the transaction.
1537 jssmag.209 initiates the next request.
1538 The `*' on the request
1539 indicates that XO (`exactly once') was \fInot\fP set.
1544 Fragmented Internet datagrams are printed as
1548 \fB(frag \fIid\fB:\fIsize\fB@\fIoffset\fB+)\fR
1549 \fB(frag \fIid\fB:\fIsize\fB@\fIoffset\fB)\fR
1553 (The first form indicates there are more fragments.
1555 indicates this is the last fragment.)
1557 \fIId\fP is the fragment id.
1558 \fISize\fP is the fragment
1559 size (in bytes) excluding the IP header.
1560 \fIOffset\fP is this
1561 fragment's offset (in bytes) in the original datagram.
1563 The fragment information is output for each fragment.
1565 fragment contains the higher level protocol header and the frag
1566 info is printed after the protocol info.
1568 after the first contain no higher level protocol header and the
1569 frag info is printed after the source and destination addresses.
1570 For example, here is part of an ftp from arizona.edu to lbl-rtsg.arpa
1571 over a CSNET connection that doesn't appear to handle 576 byte datagrams:
1575 \s-2\f(CWarizona.ftp-data > rtsg.1170: . 1024:1332(308) ack 1 win 4096 (frag 595a:328@0+)
1576 arizona > rtsg: (frag 595a:204@328)
1577 rtsg.1170 > arizona.ftp-data: . ack 1536 win 2560\fP\s+2
1581 There are a couple of things to note here: First, addresses in the
1582 2nd line don't include port numbers.
1583 This is because the TCP
1584 protocol information is all in the first fragment and we have no idea
1585 what the port or sequence numbers are when we print the later fragments.
1586 Second, the tcp sequence information in the first line is printed as if there
1587 were 308 bytes of user data when, in fact, there are 512 bytes (308 in
1588 the first frag and 204 in the second).
1589 If you are looking for holes
1590 in the sequence space or trying to match up acks
1591 with packets, this can fool you.
1593 A packet with the IP \fIdon't fragment\fP flag is marked with a
1594 trailing \fB(DF)\fP.
1598 By default, all output lines are preceded by a timestamp.
1600 is the current clock time in the form
1606 and is as accurate as the kernel's clock.
1607 The timestamp reflects the time the kernel first saw the packet.
1609 is made to account for the time lag between when the
1610 Ethernet interface removed the packet from the wire and when the kernel
1611 serviced the `new packet' interrupt.
1613 stty(1), pcap(3), bpf(4), pcap-filter(7)
1615 The original authors are:
1619 Steven McCanne, all of the
1620 Lawrence Berkeley National Laboratory, University of California, Berkeley, CA.
1622 It is currently being maintained by tcpdump.org.
1624 The current version is available via http:
1627 .I http://www.tcpdump.org/
1630 The original distribution is available via anonymous ftp:
1633 .I ftp://ftp.ee.lbl.gov/tcpdump.tar.Z
1636 IPv6/IPsec support is added by WIDE/KAME project.
1637 This program uses Eric Young's SSLeay library, under specific configurations.
1639 Please send problems, bugs, questions, desirable enhancements, patches
1643 tcpdump-workers@lists.tcpdump.org
1646 NIT doesn't let you watch your own outbound traffic, BPF will.
1647 We recommend that you use the latter.
1649 On Linux systems with 2.0[.x] kernels:
1651 packets on the loopback device will be seen twice;
1653 packet filtering cannot be done in the kernel, so that all packets must
1654 be copied from the kernel in order to be filtered in user mode;
1656 all of a packet, not just the part that's within the snapshot length,
1657 will be copied from the kernel (the 2.0[.x] packet capture mechanism, if
1658 asked to copy only part of a packet to userland, will not report the
1659 true length of the packet; this would cause most IP packets to get an
1663 capturing on some PPP devices won't work correctly.
1665 We recommend that you upgrade to a 2.2 or later kernel.
1667 Some attempt should be made to reassemble IP fragments or, at least
1668 to compute the right length for the higher level protocol.
1670 Name server inverse queries are not dumped correctly: the (empty)
1671 question section is printed rather than real query in the answer
1673 Some believe that inverse queries are themselves a bug and
1674 prefer to fix the program generating them rather than \fItcpdump\fP.
1676 A packet trace that crosses a daylight savings time change will give
1677 skewed time stamps (the time change is ignored).
1679 Filter expressions on fields other than those in Token Ring headers will
1680 not correctly handle source-routed Token Ring packets.
1682 Filter expressions on fields other than those in 802.11 headers will not
1683 correctly handle 802.11 data packets with both To DS and From DS set.
1686 should chase header chain, but at this moment it does not.
1687 .BR "ip6 protochain"
1688 is supplied for this behavior.
1690 Arithmetic expression against transport layer headers, like \fBtcp[0]\fP,
1691 does not work against IPv6 packets.
1692 It only looks at IPv4 packets.
projects / dragonfly.git / blob