1TCPDUMP(8) System Manager's Manual TCPDUMP(8)
2
6 tcpdump - dump traffic on a network
7
9 tcpdump [ -AdDefIJKlLnNOpqRStuUvxX ] [ -B buffer_size ] [ -c count ]
10 [ -C file_size ] [ -G rotate_seconds ] [ -F file ]
11 [ -i interface ] [ -j tstamp_type ] [ -m module ] [ -M secret ]
12 [ -Q|-P in|out|inout ]
13 [ -r file ] [ -s snaplen ] [ -T type ] [ -w file ]
14 [ -W filecount ]
15 [ -E spi@ipaddr algo:secret,... ]
16 [ -y datalinktype ] [ -z postrotate-command ] [ -Z user ]
17 [ expression ]
18
20 Tcpdump prints out a description of the contents of packets on a net‐
21 work interface that match the boolean expression. It can also be run
22 with the -w flag, which causes it to save the packet data to a file for
23 later analysis, and/or with the -r flag, which causes it to read from a
24 saved packet file rather than to read packets from a network interface.
25 In all cases, only packets that match expression will be processed by
26 tcpdump.
27 Tcpdump will, if not run with the -c flag, continue capturing packets
29 until it is interrupted by a SIGINT signal (generated, for example, by
30 typing your interrupt character, typically control-C) or a SIGTERM sig‐
31 nal (typically generated with the kill(1) command); if run with the -c
32 flag, it will capture packets until it is interrupted by a SIGINT or
33 SIGTERM signal or the specified number of packets have been processed.
34 When tcpdump finishes capturing packets, it will report counts of:
36 packets ``captured'' (this is the number of packets that tcpdump
38 has received and processed);
39 packets ``received by filter'' (the meaning of this depends on
41 the OS on which you're running tcpdump, and possibly on the way
42 the OS was configured - if a filter was specified on the command
43 line, on some OSes it counts packets regardless of whether they
44 were matched by the filter expression and, even if they were
45 matched by the filter expression, regardless of whether tcpdump
46 has read and processed them yet, on other OSes it counts only
47 packets that were matched by the filter expression regardless of
48 whether tcpdump has read and processed them yet, and on other
49 OSes it counts only packets that were matched by the filter
50 expression and were processed by tcpdump);
51 packets ``dropped by kernel'' (this is the number of packets
53 that were dropped, due to a lack of buffer space, by the packet
54 capture mechanism in the OS on which tcpdump is running, if the
55 OS reports that information to applications; if not, it will be
56 reported as 0).
57 On platforms that support the SIGINFO signal, such as most BSDs
59 (including Mac OS X) and Digital/Tru64 UNIX, it will report those
60 counts when it receives a SIGINFO signal (generated, for example, by
61 typing your ``status'' character, typically control-T, although on some
62 platforms, such as Mac OS X, the ``status'' character is not set by
63 default, so you must set it with stty(1) in order to use it) and will
64 continue capturing packets.
65 Reading packets from a network interface may require that you have spe‐
67 cial privileges; see the pcap (3PCAP) man page for details. Reading a
68 saved packet file doesn't require special privileges.
69
71 -A Print each packet (minus its link level header) in ASCII. Handy
72 for capturing web pages.
73 -B Set the operating system capture buffer size to buffer_size.
75 -c Exit after receiving count packets.
77 -C Before writing a raw packet to a savefile, check whether the
79 file is currently larger than file_size and, if so, close the
80 current savefile and open a new one. Savefiles after the first
81 savefile will have the name specified with the -w flag, with a
82 number after it, starting at 1 and continuing upward. The units
83 of file_size are millions of bytes (1,000,000 bytes, not
84 1,048,576 bytes).
85 Note that when used with -Z option (enabled by default), privi‐
87 leges are dropped before opening first savefile.
88 -d Dump the compiled packet-matching code in a human readable form
90 to standard output and stop.
91 -dd Dump packet-matching code as a C program fragment.
93 -ddd Dump packet-matching code as decimal numbers (preceded with a
95 count).
96 -D Print the list of the network interfaces available on the system
98 and on which tcpdump can capture packets. For each network
99 interface, a number and an interface name, possibly followed by
100 a text description of the interface, is printed. The interface
101 name or the number can be supplied to the -i flag to specify an
102 interface on which to capture.
103 This can be useful on systems that don't have a command to list
105 them (e.g., Windows systems, or UNIX systems lacking ifconfig
106 -a); the number can be useful on Windows 2000 and later systems,
107 where the interface name is a somewhat complex string.
108 The -D flag will not be supported if tcpdump was built with an
110 older version of libpcap that lacks the pcap_findalldevs() func‐
111 tion.
112 -e Print the link-level header on each dump line.
114 -E Use spi@ipaddr algo:secret for decrypting IPsec ESP packets that
116 are addressed to addr and contain Security Parameter Index value
117 spi. This combination may be repeated with comma or newline
118 seperation.
119 Note that setting the secret for IPv4 ESP packets is supported
121 at this time.
122 Algorithms may be des-cbc, 3des-cbc, blowfish-cbc, rc3-cbc,
124 cast128-cbc, or none. The default is des-cbc. The ability to
125 decrypt packets is only present if tcpdump was compiled with
126 cryptography enabled.
127 secret is the ASCII text for ESP secret key. If preceeded by
129 0x, then a hex value will be read.
130 The option assumes RFC2406 ESP, not RFC1827 ESP. The option is
132 only for debugging purposes, and the use of this option with a
133 true `secret' key is discouraged. By presenting IPsec secret
134 key onto command line you make it visible to others, via ps(1)
135 and other occasions.
136 In addition to the above syntax, the syntax file name may be
138 used to have tcpdump read the provided file in. The file is
139 opened upon receiving the first ESP packet, so any special per‐
140 missions that tcpdump may have been given should already have
141 been given up.
142 -f Print `foreign' IPv4 addresses numerically rather than symboli‐
144 cally (this option is intended to get around serious brain dam‐
145 age in Sun's NIS server — usually it hangs forever translating
146 non-local internet numbers).
147 The test for `foreign' IPv4 addresses is done using the IPv4
149 address and netmask of the interface on which capture is being
150 done. If that address or netmask are not available, available,
151 either because the interface on which capture is being done has
152 no address or netmask or because the capture is being done on
153 the Linux "any" interface, which can capture on more than one
154 interface, this option will not work correctly.
155 -F Use file as input for the filter expression. An additional
157 expression given on the command line is ignored.
158 -G If specified, rotates the dump file specified with the -w option
160 every rotate_seconds seconds. Savefiles will have the name
161 specified by -w which should include a time format as defined by
162 strftime(3). If no time format is specified, each new file will
163 overwrite the previous.
164 If used in conjunction with the -C option, filenames will take
166 the form of `file<count>'.
167 -i Listen on interface. If unspecified, tcpdump searches the sys‐
169 tem interface list for the lowest numbered, configured up inter‐
170 face (excluding loopback). Ties are broken by choosing the ear‐
171 liest match.
172 On Linux systems with 2.2 or later kernels, an interface argu‐
174 ment of ``any'' can be used to capture packets from all inter‐
175 faces. Note that captures on the ``any'' device will not be
176 done in promiscuous mode.
177 If the -D flag is supported, an interface number as printed by
179 that flag can be used as the interface argument.
180 -I Put the interface in "monitor mode"; this is supported only on
182 IEEE 802.11 Wi-Fi interfaces, and supported only on some operat‐
183 ing systems.
184 Note that in monitor mode the adapter might disassociate from
186 the network with which it's associated, so that you will not be
187 able to use any wireless networks with that adapter. This could
188 prevent accessing files on a network server, or resolving host
189 names or network addresses, if you are capturing in monitor mode
190 and are not connected to another network with another adapter.
191 This flag will affect the output of the -L flag. If -I isn't
193 specified, only those link-layer types available when not in
194 monitor mode will be shown; if -I is specified, only those link-
195 layer types available when in monitor mode will be shown.
196 -j Set the time stamp type for the capture to tstamp_type.
198 -J List the supported time stamp types for the interface and exit.
200 If the time stamp type cannot be set for the interface, no time
201 stamp types are listed.
202 --time-stamp-precision=tstamp_precision
204 When capturing, set the time stamp precision for the capture to
205 tstamp_precision. Note that availability of high precision time
206 stamps (nanoseconds) and their actual accuracy is platform and
207 hardware dependent. Also note that when writing captures made
208 with nanosecond accuracy to a savefile, the time stamps are
209 written with nanosecond resolution, and the file is written with
210 a different magic number, to indicate that the time stamps are
211 in seconds and nanoseconds; not all programs that read pcap
212 savefiles will be able to read those captures.
213 When reading a savefile, convert time stamps to the precision
215 specified by timestamp_precision, and display them with that
216 resolution. If the precision specified is less than the preci‐
217 sion of time stamps in the file, the conversion will lose preci‐
218 sion.
219 The supported values for timestamp_precision are micro for
221 microsecond resolution and nano for nanosecond resolution. The
222 default is microsecond resolution.
223 -K Don't attempt to verify IP, TCP, or UDP checksums. This is use‐
225 ful for interfaces that perform some or all of those checksum
226 calculation in hardware; otherwise, all outgoing TCP checksums
227 will be flagged as bad.
228 -l Make stdout line buffered. Useful if you want to see the data
230 while capturing it. E.g.,
231 ``tcpdump -l | tee dat'' or ``tcpdump -l >
232 dat & tail -f dat''.
233 -L List the known data link types for the interface, in the speci‐
235 fied mode, and exit. The list of known data link types may be
236 dependent on the specified mode; for example, on some platforms,
237 a Wi-Fi interface might support one set of data link types when
238 not in monitor mode (for example, it might support only fake
239 Ethernet headers, or might support 802.11 headers but not sup‐
240 port 802.11 headers with radio information) and another set of
241 data link types when in monitor mode (for example, it might sup‐
242 port 802.11 headers, or 802.11 headers with radio information,
243 only in monitor mode).
244 -m Load SMI MIB module definitions from file module. This option
246 can be used several times to load several MIB modules into tcp‐
247 dump.
248 -M Use secret as a shared secret for validating the digests found
250 in TCP segments with the TCP-MD5 option (RFC 2385), if present.
251 -n Don't convert host addresses to names. This can be used to
253 avoid DNS lookups.
254 -nn Don't convert protocol and port numbers etc. to names either.
256 -N Don't print domain name qualification of host names. E.g., if
258 you give this flag then tcpdump will print ``nic'' instead of
259 ``nic.ddn.mil''.
260 -O Do not run the packet-matching code optimizer. This is useful
262 only if you suspect a bug in the optimizer.
263 -p Don't put the interface into promiscuous mode. Note that the
265 interface might be in promiscuous mode for some other reason;
266 hence, `-p' cannot be used as an abbreviation for `ether host
267 {local-hw-addr} or ether broadcast'.
268 -Q|-P Choose send/receive direction direction for which packets should
270 be captured. Possible values are `in', `out' and `inout'. Not
271 available on all platforms.
272 -q Quick (quiet?) output. Print less protocol information so out‐
274 put lines are shorter.
275 -R Assume ESP/AH packets to be based on old specification (RFC1825
277 to RFC1829). If specified, tcpdump will not print replay pre‐
278 vention field. Since there is no protocol version field in
279 ESP/AH specification, tcpdump cannot deduce the version of
280 ESP/AH protocol.
281 -r Read packets from file (which was created with the -w option).
283 Standard input is used if file is ``-''.
284 -S Print absolute, rather than relative, TCP sequence numbers.
286 -s Snarf snaplen bytes of data from each packet rather than the
288 default of 65535 bytes. Packets truncated because of a limited
289 snapshot are indicated in the output with ``[|proto]'', where
290 proto is the name of the protocol level at which the truncation
291 has occurred. Note that taking larger snapshots both increases
292 the amount of time it takes to process packets and, effectively,
293 decreases the amount of packet buffering. This may cause pack‐
294 ets to be lost. You should limit snaplen to the smallest number
295 that will capture the protocol information you're interested in.
296 Setting snaplen to 0 sets it to the default of 65535, for back‐
297 wards compatibility with recent older versions of tcpdump.
298 -T Force packets selected by "expression" to be interpreted the
300 specified type. Currently known types are aodv (Ad-hoc On-
301 demand Distance Vector protocol), cnfp (Cisco NetFlow protocol),
302 rpc (Remote Procedure Call), rtp (Real-Time Applications proto‐
303 col), rtcp (Real-Time Applications control protocol), snmp (Sim‐
304 ple Network Management Protocol), tftp (Trivial File Transfer
305 Protocol), vat (Visual Audio Tool), and wb (distributed White
306 Board).
307 -t Don't print a timestamp on each dump line.
309 -tt Print an unformatted timestamp on each dump line.
311 -ttt Print a delta (micro-second resolution) between current and pre‐
313 vious line on each dump line.
314 -tttt Print a timestamp in default format proceeded by date on each
316 dump line.
317 -ttttt Print a delta (micro-second resolution) between current and
319 first line on each dump line.
320 -u Print undecoded NFS handles.
322 -U Make output saved via the -w option ``packet-buffered''; i.e.,
324 as each packet is saved, it will be written to the output file,
325 rather than being written only when the output buffer fills.
326 The -U flag will not be supported if tcpdump was built with an
328 older version of libpcap that lacks the pcap_dump_flush() func‐
329 tion.
330 -v When parsing and printing, produce (slightly more) verbose out‐
332 put. For example, the time to live, identification, total
333 length and options in an IP packet are printed. Also enables
334 additional packet integrity checks such as verifying the IP and
335 ICMP header checksum.
336 When writing to a file with the -w option, report, every 10 sec‐
338 onds, the number of packets captured.
339 -vv Even more verbose output. For example, additional fields are
341 printed from NFS reply packets, and SMB packets are fully
342 decoded.
343 -vvv Even more verbose output. For example, telnet SB ... SE options
345 are printed in full. With -X Telnet options are printed in hex
346 as well.
347 -w Write the raw packets to file rather than parsing and printing
349 them out. They can later be printed with the -r option. Stan‐
350 dard output is used if file is ``-''. See pcap-savefile(5) for
351 a description of the file format.
352 -W Used in conjunction with the -C option, this will limit the num‐
354 ber of files created to the specified number, and begin over‐
355 writing files from the beginning, thus creating a 'rotating'
356 buffer. In addition, it will name the files with enough leading
357 0s to support the maximum number of files, allowing them to sort
358 correctly.
359 Used in conjunction with the -G option, this will limit the num‐
361 ber of rotated dump files that get created, exiting with status
362 0 when reaching the limit. If used with -C as well, the behavior
363 will result in cyclical files per timeslice.
364 -x When parsing and printing, in addition to printing the headers
366 of each packet, print the data of each packet (minus its link
367 level header) in hex. The smaller of the entire packet or
368 snaplen bytes will be printed. Note that this is the entire
369 link-layer packet, so for link layers that pad (e.g. Ethernet),
370 the padding bytes will also be printed when the higher layer
371 packet is shorter than the required padding.
372 -xx When parsing and printing, in addition to printing the headers
374 of each packet, print the data of each packet, including its
375 link level header, in hex.
376 -X When parsing and printing, in addition to printing the headers
378 of each packet, print the data of each packet (minus its link
379 level header) in hex and ASCII. This is very handy for
380 analysing new protocols.
381 -XX When parsing and printing, in addition to printing the headers
383 of each packet, print the data of each packet, including its
384 link level header, in hex and ASCII.
385 -y Set the data link type to use while capturing packets to
387 datalinktype.
388 -z Used in conjunction with the -C or -G options, this will make
390 tcpdump run " command file " where file is the savefile being
391 closed after each rotation. For example, specifying -z gzip or
392 -z bzip2 will compress each savefile using gzip or bzip2.
393 Note that tcpdump will run the command in parallel to the cap‐
395 ture, using the lowest priority so that this doesn't disturb the
396 capture process.
397 And in case you would like to use a command that itself takes
399 flags or different arguments, you can always write a shell
400 script that will take the savefile name as the only argument,
401 make the flags & arguments arrangements and execute the command
402 that you want.
403 -Z Drops privileges (if root) and changes user ID to user and the
405 group ID to the primary group of user.
406 This behavior is enabled by default (-Z tcpdump), and can be
408 disabled by -Z root.
409 expression
412 selects which packets will be dumped. If no expression is
413 given, all packets on the net will be dumped. Otherwise, only
414 packets for which expression is `true' will be dumped.
415 For the expression syntax, see pcap-filter(7).
417 Expression arguments can be passed to tcpdump as either a single
419 argument or as multiple arguments, whichever is more convenient.
420 Generally, if the expression contains Shell metacharacters, it
421 is easier to pass it as a single, quoted argument. Multiple
422 arguments are concatenated with spaces before being parsed.
423
425 To print all packets arriving at or departing from sundown:
426 tcpdump host sundown
427 To print traffic between helios and either hot or ace:
429 tcpdump host helios and \( hot or ace \)
430 To print all IP packets between ace and any host except helios:
432 tcpdump ip host ace and not helios
433 To print all traffic between local hosts and hosts at Berkeley:
435 tcpdump net ucb-ether
436 To print all ftp traffic through internet gateway snup: (note that the
438 expression is quoted to prevent the shell from (mis-)interpreting the
439 parentheses):
440 tcpdump 'gateway snup and (port ftp or ftp-data)'
441 To print traffic neither sourced from nor destined for local hosts (if
443 you gateway to one other net, this stuff should never make it onto your
444 local net).
445 tcpdump ip and not net localnet
446 To print the start and end packets (the SYN and FIN packets) of each
448 TCP conversation that involves a non-local host.
449 tcpdump 'tcp[tcpflags] & (tcp-syn|tcp-fin) != 0 and not src and dst net localnet'
450 To print all IPv4 HTTP packets to and from port 80, i.e. print only
452 packets that contain data, not, for example, SYN and FIN packets and
453 ACK-only packets. (IPv6 is left as an exercise for the reader.)
454 tcpdump 'tcp port 80 and (((ip[2:2] - ((ip[0]&0xf)<<2)) - ((tcp[12]&0xf0)>>2)) != 0)'
455 To print IP packets longer than 576 bytes sent through gateway snup:
457 tcpdump 'gateway snup and ip[2:2] > 576'
458 To print IP broadcast or multicast packets that were not sent via Eth‐
460 ernet broadcast or multicast:
461 tcpdump 'ether[0] & 1 = 0 and ip[16] >= 224'
462 To print all ICMP packets that are not echo requests/replies (i.e., not
464 ping packets):
465 tcpdump 'icmp[icmptype] != icmp-echo and icmp[icmptype] != icmp-echoreply'
466
468 The output of tcpdump is protocol dependent. The following gives a
469 brief description and examples of most of the formats.
470 Link Level Headers
472 If the '-e' option is given, the link level header is printed out. On
474 Ethernets, the source and destination addresses, protocol, and packet
475 length are printed.
476 On FDDI networks, the '-e' option causes tcpdump to print the `frame
478 control' field, the source and destination addresses, and the packet
479 length. (The `frame control' field governs the interpretation of the
480 rest of the packet. Normal packets (such as those containing IP data‐
481 grams) are `async' packets, with a priority value between 0 and 7; for
482 example, `async4'. Such packets are assumed to contain an 802.2 Logi‐
483 cal Link Control (LLC) packet; the LLC header is printed if it is not
484 an ISO datagram or a so-called SNAP packet.
485 On Token Ring networks, the '-e' option causes tcpdump to print the
487 `access control' and `frame control' fields, the source and destination
488 addresses, and the packet length. As on FDDI networks, packets are
489 assumed to contain an LLC packet. Regardless of whether the '-e'
490 option is specified or not, the source routing information is printed
491 for source-routed packets.
492 On 802.11 networks, the '-e' option causes tcpdump to print the `frame
494 control' fields, all of the addresses in the 802.11 header, and the
495 packet length. As on FDDI networks, packets are assumed to contain an
496 LLC packet.
497 (N.B.: The following description assumes familiarity with the SLIP com‐
499 pression algorithm described in RFC-1144.)
500 On SLIP links, a direction indicator (``I'' for inbound, ``O'' for out‐
502 bound), packet type, and compression information are printed out. The
503 packet type is printed first. The three types are ip, utcp, and ctcp.
504 No further link information is printed for ip packets. For TCP pack‐
505 ets, the connection identifier is printed following the type. If the
506 packet is compressed, its encoded header is printed out. The special
507 cases are printed out as *S+n and *SA+n, where n is the amount by which
508 the sequence number (or sequence number and ack) has changed. If it is
509 not a special case, zero or more changes are printed. A change is
510 indicated by U (urgent pointer), W (window), A (ack), S (sequence num‐
511 ber), and I (packet ID), followed by a delta (+n or -n), or a new value
512 (=n). Finally, the amount of data in the packet and compressed header
513 length are printed.
514 For example, the following line shows an outbound compressed TCP
516 packet, with an implicit connection identifier; the ack has changed by
517 6, the sequence number by 49, and the packet ID by 6; there are 3 bytes
518 of data and 6 bytes of compressed header:
519 O ctcp * A+6 S+49 I+6 3 (6)
520 ARP/RARP Packets
522 Arp/rarp output shows the type of request and its arguments. The for‐
524 mat is intended to be self explanatory. Here is a short sample taken
525 from the start of an `rlogin' from host rtsg to host csam:
526 arp who-has csam tell rtsg
527 arp reply csam is-at CSAM
528 The first line says that rtsg sent an arp packet asking for the Ether‐
529 net address of internet host csam. Csam replies with its Ethernet
530 address (in this example, Ethernet addresses are in caps and internet
531 addresses in lower case).
532 This would look less redundant if we had done tcpdump -n:
534 arp who-has 128.3.254.6 tell 128.3.254.68
535 arp reply 128.3.254.6 is-at 02:07:01:00:01:c4
536 If we had done tcpdump -e, the fact that the first packet is broadcast
538 and the second is point-to-point would be visible:
539 RTSG Broadcast 0806 64: arp who-has csam tell rtsg
540 CSAM RTSG 0806 64: arp reply csam is-at CSAM
541 For the first packet this says the Ethernet source address is RTSG, the
542 destination is the Ethernet broadcast address, the type field contained
543 hex 0806 (type ETHER_ARP) and the total length was 64 bytes.
544 TCP Packets
546 (N.B.:The following description assumes familiarity with the TCP proto‐
548 col described in RFC-793. If you are not familiar with the protocol,
549 neither this description nor tcpdump will be of much use to you.)
550 The general format of a tcp protocol line is:
552 src > dst: flags data-seqno ack window urgent options
553 Src and dst are the source and destination IP addresses and ports.
554 Flags are some combination of S (SYN), F (FIN), P (PUSH), R (RST), W
555 (ECN CWR) or E (ECN-Echo), or a single `.' (no flags). Data-seqno
556 describes the portion of sequence space covered by the data in this
557 packet (see example below). Ack is sequence number of the next data
558 expected the other direction on this connection. Window is the number
559 of bytes of receive buffer space available the other direction on this
560 connection. Urg indicates there is `urgent' data in the packet.
561 Options are tcp options enclosed in angle brackets (e.g., <mss 1024>).
562 Src, dst and flags are always present. The other fields depend on the
564 contents of the packet's tcp protocol header and are output only if
565 appropriate.
566 Here is the opening portion of an rlogin from host rtsg to host csam.
568 rtsg.1023 > csam.login: S 768512:768512(0) win 4096 <mss 1024>
569 csam.login > rtsg.1023: S 947648:947648(0) ack 768513 win 4096 <mss 1024>
570 rtsg.1023 > csam.login: . ack 1 win 4096
571 rtsg.1023 > csam.login: P 1:2(1) ack 1 win 4096
572 csam.login > rtsg.1023: . ack 2 win 4096
573 rtsg.1023 > csam.login: P 2:21(19) ack 1 win 4096
574 csam.login > rtsg.1023: P 1:2(1) ack 21 win 4077
575 csam.login > rtsg.1023: P 2:3(1) ack 21 win 4077 urg 1
576 csam.login > rtsg.1023: P 3:4(1) ack 21 win 4077 urg 1
577 The first line says that tcp port 1023 on rtsg sent a packet to port
578 login on csam. The S indicates that the SYN flag was set. The packet
579 sequence number was 768512 and it contained no data. (The notation is
580 `first:last(nbytes)' which means `sequence numbers first up to but not
581 including last which is nbytes bytes of user data'.) There was no
582 piggy-backed ack, the available receive window was 4096 bytes and there
583 was a max-segment-size option requesting an mss of 1024 bytes.
584 Csam replies with a similar packet except it includes a piggy-backed
586 ack for rtsg's SYN. Rtsg then acks csam's SYN. The `.' means no flags
587 were set. The packet contained no data so there is no data sequence
588 number. Note that the ack sequence number is a small integer (1). The
589 first time tcpdump sees a tcp `conversation', it prints the sequence
590 number from the packet. On subsequent packets of the conversation, the
591 difference between the current packet's sequence number and this ini‐
592 tial sequence number is printed. This means that sequence numbers
593 after the first can be interpreted as relative byte positions in the
594 conversation's data stream (with the first data byte each direction
595 being `1'). `-S' will override this feature, causing the original
596 sequence numbers to be output.
597 On the 6th line, rtsg sends csam 19 bytes of data (bytes 2 through 20
599 in the rtsg → csam side of the conversation). The PUSH flag is set in
600 the packet. On the 7th line, csam says it's received data sent by rtsg
601 up to but not including byte 21. Most of this data is apparently sit‐
602 ting in the socket buffer since csam's receive window has gotten 19
603 bytes smaller. Csam also sends one byte of data to rtsg in this
604 packet. On the 8th and 9th lines, csam sends two bytes of urgent,
605 pushed data to rtsg.
606 If the snapshot was small enough that tcpdump didn't capture the full
608 TCP header, it interprets as much of the header as it can and then
609 reports ``[|tcp]'' to indicate the remainder could not be interpreted.
610 If the header contains a bogus option (one with a length that's either
611 too small or beyond the end of the header), tcpdump reports it as
612 ``[bad opt]'' and does not interpret any further options (since it's
613 impossible to tell where they start). If the header length indicates
614 options are present but the IP datagram length is not long enough for
615 the options to actually be there, tcpdump reports it as ``[bad hdr
616 length]''.
617 Capturing TCP packets with particular flag combinations (SYN-ACK, URG-
619 ACK, etc.)
620 There are 8 bits in the control bits section of the TCP header:
622 CWR | ECE | URG | ACK | PSH | RST | SYN | FIN
624 Let's assume that we want to watch packets used in establishing a TCP
626 connection. Recall that TCP uses a 3-way handshake protocol when it
627 initializes a new connection; the connection sequence with regard to
628 the TCP control bits is
629 1) Caller sends SYN
631 2) Recipient responds with SYN, ACK
632 3) Caller sends ACK
633 Now we're interested in capturing packets that have only the SYN bit
635 set (Step 1). Note that we don't want packets from step 2 (SYN-ACK),
636 just a plain initial SYN. What we need is a correct filter expression
637 for tcpdump.
638 Recall the structure of a TCP header without options:
640 0 15 31
642 -----------------------------------------------------------------
643 | source port | destination port |
644 -----------------------------------------------------------------
645 | sequence number |
646 -----------------------------------------------------------------
647 | acknowledgment number |
648 -----------------------------------------------------------------
649 | HL | rsvd |C|E|U|A|P|R|S|F| window size |
650 -----------------------------------------------------------------
651 | TCP checksum | urgent pointer |
652 -----------------------------------------------------------------
653 A TCP header usually holds 20 octets of data, unless options are
655 present. The first line of the graph contains octets 0 - 3, the second
656 line shows octets 4 - 7 etc.
657 Starting to count with 0, the relevant TCP control bits are contained
659 in octet 13:
660 0 7| 15| 23| 31
662 ----------------|---------------|---------------|----------------
663 | HL | rsvd |C|E|U|A|P|R|S|F| window size |
664 ----------------|---------------|---------------|----------------
665 | | 13th octet | | |
666 Let's have a closer look at octet no. 13:
668 | |
670 |---------------|
671 |C|E|U|A|P|R|S|F|
672 |---------------|
673 |7 5 3 0|
674 These are the TCP control bits we are interested in. We have numbered
676 the bits in this octet from 0 to 7, right to left, so the PSH bit is
677 bit number 3, while the URG bit is number 5.
678 Recall that we want to capture packets with only SYN set. Let's see
680 what happens to octet 13 if a TCP datagram arrives with the SYN bit set
681 in its header:
682 |C|E|U|A|P|R|S|F|
684 |---------------|
685 |0 0 0 0 0 0 1 0|
686 |---------------|
687 |7 6 5 4 3 2 1 0|
688 Looking at the control bits section we see that only bit number 1 (SYN)
690 is set.
691 Assuming that octet number 13 is an 8-bit unsigned integer in network
693 byte order, the binary value of this octet is
694 00000010
696 and its decimal representation is
698 7 6 5 4 3 2 1 0
700 0*2 + 0*2 + 0*2 + 0*2 + 0*2 + 0*2 + 1*2 + 0*2 = 2
701 We're almost done, because now we know that if only SYN is set, the
703 value of the 13th octet in the TCP header, when interpreted as a 8-bit
704 unsigned integer in network byte order, must be exactly 2.
705 This relationship can be expressed as
707 tcp[13] == 2
708 We can use this expression as the filter for tcpdump in order to watch
710 packets which have only SYN set:
711 tcpdump -i xl0 tcp[13] == 2
712 The expression says "let the 13th octet of a TCP datagram have the dec‐
714 imal value 2", which is exactly what we want.
715 Now, let's assume that we need to capture SYN packets, but we don't
717 care if ACK or any other TCP control bit is set at the same time.
718 Let's see what happens to octet 13 when a TCP datagram with SYN-ACK set
719 arrives:
720 |C|E|U|A|P|R|S|F|
722 |---------------|
723 |0 0 0 1 0 0 1 0|
724 |---------------|
725 |7 6 5 4 3 2 1 0|
726 Now bits 1 and 4 are set in the 13th octet. The binary value of octet
728 13 is
729 00010010
731 which translates to decimal
733 7 6 5 4 3 2 1 0
735 0*2 + 0*2 + 0*2 + 1*2 + 0*2 + 0*2 + 1*2 + 0*2 = 18
736 Now we can't just use 'tcp[13] == 18' in the tcpdump filter expression,
738 because that would select only those packets that have SYN-ACK set, but
739 not those with only SYN set. Remember that we don't care if ACK or any
740 other control bit is set as long as SYN is set.
741 In order to achieve our goal, we need to logically AND the binary value
743 of octet 13 with some other value to preserve the SYN bit. We know
744 that we want SYN to be set in any case, so we'll logically AND the
745 value in the 13th octet with the binary value of a SYN:
746 00010010 SYN-ACK 00000010 SYN
748 AND 00000010 (we want SYN) AND 00000010 (we want SYN)
749 -------- --------
750 = 00000010 = 00000010
751 We see that this AND operation delivers the same result regardless
753 whether ACK or another TCP control bit is set. The decimal representa‐
754 tion of the AND value as well as the result of this operation is 2
755 (binary 00000010), so we know that for packets with SYN set the follow‐
756 ing relation must hold true:
757 ( ( value of octet 13 ) AND ( 2 ) ) == ( 2 )
759 This points us to the tcpdump filter expression
761 tcpdump -i xl0 'tcp[13] & 2 == 2'
762 Note that you should use single quotes or a backslash in the expression
764 to hide the AND ('&') special character from the shell.
765 UDP Packets
767 UDP format is illustrated by this rwho packet:
769 actinide.who > broadcast.who: udp 84
770 This says that port who on host actinide sent a udp datagram to port
771 who on host broadcast, the Internet broadcast address. The packet con‐
772 tained 84 bytes of user data.
773 Some UDP services are recognized (from the source or destination port
775 number) and the higher level protocol information printed. In particu‐
776 lar, Domain Name service requests (RFC-1034/1035) and Sun RPC calls
777 (RFC-1050) to NFS.
778 UDP Name Server Requests
780 (N.B.:The following description assumes familiarity with the Domain
782 Service protocol described in RFC-1035. If you are not familiar with
783 the protocol, the following description will appear to be written in
784 greek.)
785 Name server requests are formatted as
787 src > dst: id op? flags qtype qclass name (len)
788 h2opolo.1538 > helios.domain: 3+ A? ucbvax.berkeley.edu. (37)
789 Host h2opolo asked the domain server on helios for an address record
790 (qtype=A) associated with the name ucbvax.berkeley.edu. The query id
791 was `3'. The `+' indicates the recursion desired flag was set. The
792 query length was 37 bytes, not including the UDP and IP protocol head‐
793 ers. The query operation was the normal one, Query, so the op field
794 was omitted. If the op had been anything else, it would have been
795 printed between the `3' and the `+'. Similarly, the qclass was the
796 normal one, C_IN, and omitted. Any other qclass would have been
797 printed immediately after the `A'.
798 A few anomalies are checked and may result in extra fields enclosed in
800 square brackets: If a query contains an answer, authority records or
801 additional records section, ancount, nscount, or arcount are printed as
802 `[na]', `[nn]' or `[nau]' where n is the appropriate count. If any of
803 the response bits are set (AA, RA or rcode) or any of the `must be
804 zero' bits are set in bytes two and three, `[b2&3=x]' is printed, where
805 x is the hex value of header bytes two and three.
806 UDP Name Server Responses
808 Name server responses are formatted as
810 src > dst: id op rcode flags a/n/au type class data (len)
811 helios.domain > h2opolo.1538: 3 3/3/7 A 128.32.137.3 (273)
812 helios.domain > h2opolo.1537: 2 NXDomain* 0/1/0 (97)
813 In the first example, helios responds to query id 3 from h2opolo with 3
814 answer records, 3 name server records and 7 additional records. The
815 first answer record is type A (address) and its data is internet
816 address 128.32.137.3. The total size of the response was 273 bytes,
817 excluding UDP and IP headers. The op (Query) and response code (NoEr‐
818 ror) were omitted, as was the class (C_IN) of the A record.
819 In the second example, helios responds to query 2 with a response code
821 of non-existent domain (NXDomain) with no answers, one name server and
822 no authority records. The `*' indicates that the authoritative answer
823 bit was set. Since there were no answers, no type, class or data were
824 printed.
825 Other flag characters that might appear are `-' (recursion available,
827 RA, not set) and `|' (truncated message, TC, set). If the `question'
828 section doesn't contain exactly one entry, `[nq]' is printed.
829 SMB/CIFS decoding
832 tcpdump now includes fairly extensive SMB/CIFS/NBT decoding for data on
834 UDP/137, UDP/138 and TCP/139. Some primitive decoding of IPX and Net‐
835 BEUI SMB data is also done.
836 By default a fairly minimal decode is done, with a much more detailed
838 decode done if -v is used. Be warned that with -v a single SMB packet
839 may take up a page or more, so only use -v if you really want all the
840 gory details.
841 For information on SMB packet formats and what all te fields mean see
843 www.cifs.org or the pub/samba/specs/ directory on your favorite
844 samba.org mirror site. The SMB patches were written by Andrew Tridgell
845 (tridge@samba.org).
846 NFS Requests and Replies
849 Sun NFS (Network File System) requests and replies are printed as:
851 src.xid > dst.nfs: len op args
852 src.nfs > dst.xid: reply stat len op results
853 sushi.6709 > wrl.nfs: 112 readlink fh 21,24/10.73165
854 wrl.nfs > sushi.6709: reply ok 40 readlink "../var"
855 sushi.201b > wrl.nfs:
856 144 lookup fh 9,74/4096.6878 "xcolors"
857 wrl.nfs > sushi.201b:
858 reply ok 128 lookup fh 9,74/4134.3150
859 In the first line, host sushi sends a transaction with id 6709 to wrl
860 (note that the number following the src host is a transaction id, not
861 the source port). The request was 112 bytes, excluding the UDP and IP
862 headers. The operation was a readlink (read symbolic link) on file
863 handle (fh) 21,24/10.731657119. (If one is lucky, as in this case, the
864 file handle can be interpreted as a major,minor device number pair,
865 followed by the inode number and generation number.) Wrl replies `ok'
866 with the contents of the link.
867 In the third line, sushi asks wrl to lookup the name `xcolors' in
869 directory file 9,74/4096.6878. Note that the data printed depends on
870 the operation type. The format is intended to be self explanatory if
871 read in conjunction with an NFS protocol spec.
872 If the -v (verbose) flag is given, additional information is printed.
874 For example:
875 sushi.1372a > wrl.nfs:
876 148 read fh 21,11/12.195 8192 bytes @ 24576
877 wrl.nfs > sushi.1372a:
878 reply ok 1472 read REG 100664 ids 417/0 sz 29388
879 (-v also prints the IP header TTL, ID, length, and fragmentation
880 fields, which have been omitted from this example.) In the first line,
881 sushi asks wrl to read 8192 bytes from file 21,11/12.195, at byte off‐
882 set 24576. Wrl replies `ok'; the packet shown on the second line is
883 the first fragment of the reply, and hence is only 1472 bytes long (the
884 other bytes will follow in subsequent fragments, but these fragments do
885 not have NFS or even UDP headers and so might not be printed, depending
886 on the filter expression used). Because the -v flag is given, some of
887 the file attributes (which are returned in addition to the file data)
888 are printed: the file type (``REG'', for regular file), the file mode
889 (in octal), the uid and gid, and the file size.
890 If the -v flag is given more than once, even more details are printed.
892 Note that NFS requests are very large and much of the detail won't be
894 printed unless snaplen is increased. Try using `-s 192' to watch NFS
895 traffic.
896 NFS reply packets do not explicitly identify the RPC operation.
898 Instead, tcpdump keeps track of ``recent'' requests, and matches them
899 to the replies using the transaction ID. If a reply does not closely
900 follow the corresponding request, it might not be parsable.
901 AFS Requests and Replies
903 Transarc AFS (Andrew File System) requests and replies are printed as:
905 src.sport > dst.dport: rx packet-type
907 src.sport > dst.dport: rx packet-type service call call-name args
908 src.sport > dst.dport: rx packet-type service reply call-name args
909 elvis.7001 > pike.afsfs:
910 rx data fs call rename old fid 536876964/1/1 ".newsrc.new"
911 new fid 536876964/1/1 ".newsrc"
912 pike.afsfs > elvis.7001: rx data fs reply rename
913 In the first line, host elvis sends a RX packet to pike. This was a RX
914 data packet to the fs (fileserver) service, and is the start of an RPC
915 call. The RPC call was a rename, with the old directory file id of
916 536876964/1/1 and an old filename of `.newsrc.new', and a new directory
917 file id of 536876964/1/1 and a new filename of `.newsrc'. The host
918 pike responds with a RPC reply to the rename call (which was success‐
919 ful, because it was a data packet and not an abort packet).
920 In general, all AFS RPCs are decoded at least by RPC call name. Most
922 AFS RPCs have at least some of the arguments decoded (generally only
923 the `interesting' arguments, for some definition of interesting).
924 The format is intended to be self-describing, but it will probably not
926 be useful to people who are not familiar with the workings of AFS and
927 RX.
928 If the -v (verbose) flag is given twice, acknowledgement packets and
930 additional header information is printed, such as the the RX call ID,
931 call number, sequence number, serial number, and the RX packet flags.
932 If the -v flag is given twice, additional information is printed, such
934 as the the RX call ID, serial number, and the RX packet flags. The MTU
935 negotiation information is also printed from RX ack packets.
936 If the -v flag is given three times, the security index and service id
938 are printed.
939 Error codes are printed for abort packets, with the exception of Ubik
941 beacon packets (because abort packets are used to signify a yes vote
942 for the Ubik protocol).
943 Note that AFS requests are very large and many of the arguments won't
945 be printed unless snaplen is increased. Try using `-s 256' to watch
946 AFS traffic.
947 AFS reply packets do not explicitly identify the RPC operation.
949 Instead, tcpdump keeps track of ``recent'' requests, and matches them
950 to the replies using the call number and service ID. If a reply does
951 not closely follow the corresponding request, it might not be parsable.
952 KIP AppleTalk (DDP in UDP)
955 AppleTalk DDP packets encapsulated in UDP datagrams are de-encapsulated
957 and dumped as DDP packets (i.e., all the UDP header information is dis‐
958 carded). The file /etc/atalk.names is used to translate AppleTalk net
959 and node numbers to names. Lines in this file have the form
960 number name
961 1.254 ether
963 16.1 icsd-net
964 1.254.110 ace
965 The first two lines give the names of AppleTalk networks. The third
966 line gives the name of a particular host (a host is distinguished from
967 a net by the 3rd octet in the number - a net number must have two
968 octets and a host number must have three octets.) The number and name
969 should be separated by whitespace (blanks or tabs). The
970 /etc/atalk.names file may contain blank lines or comment lines (lines
971 starting with a `#').
972 AppleTalk addresses are printed in the form
974 net.host.port
975 144.1.209.2 > icsd-net.112.220
977 office.2 > icsd-net.112.220
978 jssmag.149.235 > icsd-net.2
979 (If the /etc/atalk.names doesn't exist or doesn't contain an entry for
980 some AppleTalk host/net number, addresses are printed in numeric form.)
981 In the first example, NBP (DDP port 2) on net 144.1 node 209 is sending
982 to whatever is listening on port 220 of net icsd node 112. The second
983 line is the same except the full name of the source node is known
984 (`office'). The third line is a send from port 235 on net jssmag node
985 149 to broadcast on the icsd-net NBP port (note that the broadcast
986 address (255) is indicated by a net name with no host number - for this
987 reason it's a good idea to keep node names and net names distinct in
988 /etc/atalk.names).
989 NBP (name binding protocol) and ATP (AppleTalk transaction protocol)
991 packets have their contents interpreted. Other protocols just dump the
992 protocol name (or number if no name is registered for the protocol) and
993 packet size.
994 NBP packets are formatted like the following examples:
996 icsd-net.112.220 > jssmag.2: nbp-lkup 190: "=:LaserWriter@*"
997 jssmag.209.2 > icsd-net.112.220: nbp-reply 190: "RM1140:LaserWriter@*" 250
998 techpit.2 > icsd-net.112.220: nbp-reply 190: "techpit:LaserWriter@*" 186
999 The first line is a name lookup request for laserwriters sent by net
1000 icsd host 112 and broadcast on net jssmag. The nbp id for the lookup
1001 is 190. The second line shows a reply for this request (note that it
1002 has the same id) from host jssmag.209 saying that it has a laserwriter
1003 resource named "RM1140" registered on port 250. The third line is
1004 another reply to the same request saying host techpit has laserwriter
1005 "techpit" registered on port 186.
1006 ATP packet formatting is demonstrated by the following example:
1008 jssmag.209.165 > helios.132: atp-req 12266<0-7> 0xae030001
1009 helios.132 > jssmag.209.165: atp-resp 12266:0 (512) 0xae040000
1010 helios.132 > jssmag.209.165: atp-resp 12266:1 (512) 0xae040000
1011 helios.132 > jssmag.209.165: atp-resp 12266:2 (512) 0xae040000
1012 helios.132 > jssmag.209.165: atp-resp 12266:3 (512) 0xae040000
1013 helios.132 > jssmag.209.165: atp-resp 12266:4 (512) 0xae040000
1014 helios.132 > jssmag.209.165: atp-resp 12266:5 (512) 0xae040000
1015 helios.132 > jssmag.209.165: atp-resp 12266:6 (512) 0xae040000
1016 helios.132 > jssmag.209.165: atp-resp*12266:7 (512) 0xae040000
1017 jssmag.209.165 > helios.132: atp-req 12266<3,5> 0xae030001
1018 helios.132 > jssmag.209.165: atp-resp 12266:3 (512) 0xae040000
1019 helios.132 > jssmag.209.165: atp-resp 12266:5 (512) 0xae040000
1020 jssmag.209.165 > helios.132: atp-rel 12266<0-7> 0xae030001
1021 jssmag.209.133 > helios.132: atp-req* 12267<0-7> 0xae030002
1022 Jssmag.209 initiates transaction id 12266 with host helios by request‐
1023 ing up to 8 packets (the `<0-7>'). The hex number at the end of the
1024 line is the value of the `userdata' field in the request.
1025 Helios responds with 8 512-byte packets. The `:digit' following the
1027 transaction id gives the packet sequence number in the transaction and
1028 the number in parens is the amount of data in the packet, excluding the
1029 atp header. The `*' on packet 7 indicates that the EOM bit was set.
1030 Jssmag.209 then requests that packets 3 & 5 be retransmitted. Helios
1032 resends them then jssmag.209 releases the transaction. Finally, jss‐
1033 mag.209 initiates the next request. The `*' on the request indicates
1034 that XO (`exactly once') was not set.
1035 IP Fragmentation
1038 Fragmented Internet datagrams are printed as
1040 (frag id:size@offset+)
1041 (frag id:size@offset)
1042 (The first form indicates there are more fragments. The second indi‐
1043 cates this is the last fragment.)
1044 Id is the fragment id. Size is the fragment size (in bytes) excluding
1046 the IP header. Offset is this fragment's offset (in bytes) in the
1047 original datagram.
1048 The fragment information is output for each fragment. The first frag‐
1050 ment contains the higher level protocol header and the frag info is
1051 printed after the protocol info. Fragments after the first contain no
1052 higher level protocol header and the frag info is printed after the
1053 source and destination addresses. For example, here is part of an ftp
1054 from arizona.edu to lbl-rtsg.arpa over a CSNET connection that doesn't
1055 appear to handle 576 byte datagrams:
1056 arizona.ftp-data > rtsg.1170: . 1024:1332(308) ack 1 win 4096 (frag 595a:328@0+)
1057 arizona > rtsg: (frag 595a:204@328)
1058 rtsg.1170 > arizona.ftp-data: . ack 1536 win 2560
1059 There are a couple of things to note here: First, addresses in the 2nd
1060 line don't include port numbers. This is because the TCP protocol
1061 information is all in the first fragment and we have no idea what the
1062 port or sequence numbers are when we print the later fragments. Sec‐
1063 ond, the tcp sequence information in the first line is printed as if
1064 there were 308 bytes of user data when, in fact, there are 512 bytes
1065 (308 in the first frag and 204 in the second). If you are looking for
1066 holes in the sequence space or trying to match up acks with packets,
1067 this can fool you.
1068 A packet with the IP don't fragment flag is marked with a trailing
1070 (DF).
1071 Timestamps
1073 By default, all output lines are preceded by a timestamp. The time‐
1075 stamp is the current clock time in the form
1076 hh:mm:ss.frac
1077 and is as accurate as the kernel's clock. The timestamp reflects the
1078 time the kernel first saw the packet. No attempt is made to account
1079 for the time lag between when the Ethernet interface removed the packet
1080 from the wire and when the kernel serviced the `new packet' interrupt.
1081
1083 stty(1), pcap(3PCAP), bpf(4), nit(4P), pcap-savefile(5), pcap-filter(7)
1084
1086 The original authors are:
1087 Van Jacobson, Craig Leres and Steven McCanne, all of the Lawrence
1089 Berkeley National Laboratory, University of California, Berkeley, CA.
1090 It is currently being maintained by tcpdump.org.
1092 The current version is available via http:
1094 http://www.tcpdump.org/
1096 The original distribution is available via anonymous ftp:
1098 ftp://ftp.ee.lbl.gov/tcpdump.tar.Z
1100 IPv6/IPsec support is added by WIDE/KAME project. This program uses
1102 Eric Young's SSLeay library, under specific configurations.
1103
1105 Please send problems, bugs, questions, desirable enhancements, patches
1106 etc. to:
1107 tcpdump-workers@lists.tcpdump.org
1109 NIT doesn't let you watch your own outbound traffic, BPF will. We rec‐
1111 ommend that you use the latter.
1112 On Linux systems with 2.0[.x] kernels:
1114 packets on the loopback device will be seen twice;
1116 packet filtering cannot be done in the kernel, so that all pack‐
1118 ets must be copied from the kernel in order to be filtered in
1119 user mode;
1120 all of a packet, not just the part that's within the snapshot
1122 length, will be copied from the kernel (the 2.0[.x] packet cap‐
1123 ture mechanism, if asked to copy only part of a packet to user‐
1124 land, will not report the true length of the packet; this would
1125 cause most IP packets to get an error from tcpdump);
1126 capturing on some PPP devices won't work correctly.
1128 We recommend that you upgrade to a 2.2 or later kernel.
1130 Some attempt should be made to reassemble IP fragments or, at least to
1132 compute the right length for the higher level protocol.
1133 Name server inverse queries are not dumped correctly: the (empty) ques‐
1135 tion section is printed rather than real query in the answer section.
1136 Some believe that inverse queries are themselves a bug and prefer to
1137 fix the program generating them rather than tcpdump.
1138 A packet trace that crosses a daylight savings time change will give
1140 skewed time stamps (the time change is ignored).
1141 Filter expressions on fields other than those in Token Ring headers
1143 will not correctly handle source-routed Token Ring packets.
1144 Filter expressions on fields other than those in 802.11 headers will
1146 not correctly handle 802.11 data packets with both To DS and From DS
1147 set.
1148 ip6 proto should chase header chain, but at this moment it does not.
1150 ip6 protochain is supplied for this behavior.
1151 Arithmetic expression against transport layer headers, like tcp[0],
1153 does not work against IPv6 packets. It only looks at IPv4 packets.
1154 05 March 2009 TCPDUMP(8)