4 tethereal - Dump and analyze network traffic
9 S<[ B<-a> capture autostop condition ] ...>
10 S<[ B<-b> number of ring buffer files ]>
13 S<[ B<-f> capture filter expression ]>
14 S<[ B<-F> file format ]>
16 S<[ B<-i> interface ]>
19 S<[ B<-N> resolving flags ]>
20 S<[ B<-o> preference setting ] ...>
24 S<[ B<-R> display filter expression ]>
27 S<[ B<-t> time stamp format ]>
32 S<[ B<-z> statistics-string ]>
33 S<[ filter expression ]>
37 B<Tethereal> is a network protocol analyzer. It lets you capture packet
38 data from a live network, or read packets from a previously saved
39 capture file, either printing a decoded form of those packets to the
40 standard output or writing the packets to a file. B<Tethereal>'s native
41 capture file format is B<libpcap> format, which is also the format used
42 by B<tcpdump> and various other tools. In addition, B<Tethereal> can
43 read capture files from B<snoop> and B<atmsnoop>, Shomiti/Finisar
44 B<Surveyor>, Novell B<LANalyzer>, Network General/Network Associates
45 DOS-based B<Sniffer> (compressed or uncompressed), Microsoft B<Network
46 Monitor>, AIX's B<iptrace>, Cinco Networks B<NetXRay>, Network
47 Associates Windows-based B<Sniffer>, AG Group/WildPackets
48 B<EtherPeek>/B<TokenPeek>/B<AiroPeek>, B<RADCOM>'s WAN/LAN analyzer,
49 B<Lucent/Ascend> router debug output, HP-UX's B<nettl>, the dump output
50 from B<Toshiba's> ISDN routers, the output from B<i4btrace> from the
51 ISDN4BSD project, the output in B<IPLog> format from the Cisco Secure
52 Intrusion Detection System, B<pppd logs> (pppdump format), the output
53 from VMS's B<TCPIPtrace> utility, the text output from the B<DBS
54 Etherwatch> VMS utility, traffic capture files from Visual Networks'
55 Visual UpTime, and the output from B<CoSine> L2 debug. There is no
56 need to tell B<Tethereal> what type of file you are reading; it will
57 determine the file type by itself. B<Tethereal> is also capable of
58 reading any of these file formats if they are compressed using gzip.
59 B<Tethereal> recognizes this directly from the file; the '.gz' extension
60 is not required for this purpose.
62 If the B<-w> flag is not specified, B<Tethereal> prints a decoded form
63 of the packets it captures or reads; otherwise, it writes those packets
64 to the file specified by that flag.
66 When printing a decoded form of packets, B<Tethereal> prints, by
67 default, a summary line containing the fields specified by the
68 preferences file (which are also the fields displayed in the packet list
69 pane in B<Ethereal>), although if it's printing packets as it captures
70 them, rather than printing packets from a saved capture file, it won't
71 print the "frame number" field. If the B<-V> flag is specified, it
72 prints intead a protocol tree, showing all the fields of all protocols
75 When writing packets to a file, B<Tethereal>, by default, writes the
76 file in B<libpcap> format, and writes all of the packets it sees to the
77 output file. The B<-F> flag can be used to specify the format in which
78 to write the file; it can write the file in B<libpcap> format (standard
79 B<libpcap> format, a modified format used by some patched versions of
80 B<libpcap>, or the format used by Red Hat Linux 6.1), B<snoop> format,
81 uncompressed B<Sniffer> format, Microsoft B<Network Monitor> 1.x format,
82 the format used by Windows-based versions of the B<Sniffer>
83 software, and the format used by Visual Networks' software.
85 Read filters in B<Tethereal>, which allow you to select which packets
86 are to be decoded or written to a file, are very powerful; more fields
87 are filterable in B<Tethereal> than in other protocol analyzers, and the
88 syntax you can use to create your filters is richer. As B<Tethereal>
89 progresses, expect more and more protocol fields to be allowed in read
92 Packet capturing is performed with the pcap library. The capture filter
93 syntax follows the rules of the pcap library. This syntax is different
94 from the read filter syntax. A read filter can also be specified when
95 capturing, and only packets that pass the read filter will be displayed
96 or saved to the output file; note, however, that capture filters are much
97 more efficient than read filters, and it may be more difficult for
98 B<Tethereal> to keep up with a busy network if a read filter is
99 specified for a live capture.
101 Compressed file support uses (and therefore requires) the zlib library.
102 If the zlib library is not present, B<Tethereal> will compile, but will
103 be unable to read compressed files.
105 A capture or read filter can either be specified with the B<-f> or B<-R>
106 option, respectively, in which case the entire filter expression must be
107 specified as a single argument (which means that if it contains spaces,
108 it must be quoted), or can be specified with command-line arguments
109 after the option arguments, in which case all the arguments after the
110 filter arguments are treated as a filter expression. If the filter is
111 specified with command-line arguments after the option arguments, it's a
112 capture filter if a capture is being done (i.e., if no B<-r> flag was
113 specified) and a read filter if a capture file is being read (i.e., if a
114 B<-r> flag was specified).
122 Specify a criterion that specifies when B<Tethereal> is to stop writing
123 to a capture file. The criterion is of the form I<test>B<:>I<value>,
124 where I<test> is one of:
132 Stop writing to a capture file after I<value> seconds have elapsed.
136 Stop writing to a capture file after it reaches a size of I<value>
137 kilobytes (where a kilobyte is 1000 bytes, not 1024 bytes).
145 If a maximum capture file size was specified, cause B<Tethereal> to run
146 in "ring buffer" mode, with the specified number of files. In "ring
147 buffer" mode, B<Tethereal> will write to several capture files; the name
148 of the first file, while the capture is in progress, will be the name
149 specified by the B<-w> flag, and subsequent files with have .I<n>
150 appended, with I<n> counting up.
152 When the first capture file fills up, B<Tethereal> will switch to
153 writing to the next file, until it fills up the last file, at which
154 point it'll discard the data in the first file and start writing to that
155 file. When that file fills up, B<Tethereal> will discard the data in
156 the next file and start writing to it, and so on.
158 When the capture completes, the files will be renamed to have names
159 based on the number of the file and on the date and time at which
160 packets most recently started being written to the file.
162 You can only save files in B<libpcap> format when using a ring buffer.
166 Set the default number of packets to read when capturing live
171 Print a list of the interfaces on which B<Tethereal> can capture, and
172 exit. Note that "can capture" means that B<Tethereal> was able to open
173 that device to do a live capture; if, on your system, a program doing a
174 network capture must be run from an account with special privileges (for
175 example, as root), then, if B<Tethereal> is run with the B<-D> flag and
176 is not run from such an account, it will not list any interfaces.
180 Set the capture filter expression.
184 Set the file format of the output capture file.
188 Print the version and options and exits.
192 Set the name of the network interface to use for live packet capture.
193 It should match one of the names listed in "B<netstat -i>" or
194 "B<ifconfig -a>". If no interface is specified, B<Tethereal> searches
195 the list of interfaces, choosing the first non-loopback interface if
196 there are any non-loopback interfaces, and choosing the first loopback
197 interface if there are no non-loopback interfaces; if there are no
198 interfaces, B<Tethereal> reports an error and doesn't start the capture.
202 Flush the standard output after the information for each packet is
203 printed. (This is not, strictly speaking, line-buffered if B<-V>
204 was specified; however, it is the same as line-buffered if B<-V> wasn't
205 specified, as only one line is printed for each packet, and, as B<-l> is
206 normally used when piping a live capture to a program or script, so that
207 output for a packet shows up as soon as the packet is seen and
208 dissected, it should work just as well as true line-buffering. We do
209 this as a workaround for a deficiency in the Microsoft Visual C++ C
212 This may be useful when piping the output of B<Tethereal> to another
213 program, as it means that the program to which the output is piped will
214 see the dissected data for a packet as soon as B<Tethereal> sees the
215 packet and generates that output, rather than seeing it only when the
216 standard output buffer containing that data fills up.
220 Disable network object name resolution (such as hostname, TCP and UDP port
225 Turn on name resolving for particular types of addresses and port
226 numbers, with name resolving for other types of addresses and port
227 numbers turned off; the argument is a string that may contain the
228 letters B<m> to enable MAC address resolution, B<n> to enable network
229 address resolution, and B<t> to enable transport-layer port number
230 resolution. This overrides B<-n> if both B<-N> and B<-n> are present.
234 Set a preference value, overriding the default value and any value read
235 from a preference file. The argument to the flag is a string of the
236 form I<prefname>B<:>I<value>, where I<prefname> is the name of the
237 preference (which is the same name that would appear in the preference
238 file), and I<value> is the value to which it should be set.
242 I<Don't> put the interface into promiscuous mode. Note that the
243 interface might be in promiscuous mode for some other reason; hence,
244 B<-p> cannot be used to ensure that the only traffic that is captured is
245 traffic sent to or from the machine on which B<Tethereal> is running,
246 broadcast traffic, and multicast traffic to addresses received by that
251 Don't display the continuous count of packets captured that is normally
252 shown when saving a capture to a file; instead, just display, at the end
253 of the capture, a count of packets captured.
257 Read packet data from I<infile>.
261 Cause the specified filter (which uses the syntax of read filters,
262 rather than that of capture filters) to be applied before printing a
263 decoded form of packets or writing packets to a file; packets not
264 matching the filter are discarded rather than being printed or written.
268 Set the default snapshot length to use when capturing live data.
269 No more than I<snaplen> bytes of each network packet will be read into
270 memory, or saved to disk.
274 Decode and display packets even while writing to file.
278 Set the format of the packet timestamp printed in summary lines. The
279 format can be one of 'r' (relative), 'a' (absolute), 'ad' (absolute with
280 date), or 'd' (delta). The relative time is the time elapsed between
281 the first packet and the current packet. The absolute time is the
282 actual time the packet was captured, with no date displayed; the
283 absolute date and time is the actual time and date the packet was
284 captured. The delta time is the time since the previous packet was
285 captured. The default is relative.
289 Print the version and exit.
293 Cause B<Tethereal> to print a protocol tree for each packet rather than
294 a one-line summary of the packet.
298 Write packet data to I<savefile> or to the standard output if
303 Cause B<Tethereal> to print a hex and ASCII dump of the packet data
304 after printing the summary or protocol tree.
308 Get B<Tethereal> to collect various types of statistics and display the result
309 after finishing reading the capture file.
310 Currently implemented statistics are:
312 B<-z> rpc,rtt,I<program>,I<version>
314 Collect call/reply RTT data for I<program>/I<version>. Data collected
315 is number of calls for each procedure, MinRTT, MaxRTT and AvgRTT.
316 Example: use B<-z rpc,rtt,100003,3> to collect data for NFS v3. This
317 option can be used multiple times on the command line.
321 Collect call/reply RTT data for all known ONC-RPC programs/versions.
322 Data collected is number of calls for each protocol/version, MinRTT,
324 This option can only be used once on the command line.
328 =head1 CAPTURE FILTER SYNTAX
330 See manual page of tcpdump(8).
332 =head1 READ FILTER SYNTAX
334 Read filters help you remove the noise from a packet trace and let you
335 see only the packets that interest you. If a packet meets the
336 requirements expressed in your read filter, then it is printed. Read
337 filters let you compare the fields within a protocol against a specific
338 value, compare fields against fields, and to check the existence of
339 specified fields or protocols.
341 The simplest read filter allows you to check for the existence of a
342 protocol or field. If you want to see all packets which contain the IPX
343 protocol, the filter would be "ipx". (Without the quotation marks) To
344 see all packets that contain a Token-Ring RIF field, use "tr.rif".
346 Fields can also be compared against values. The comparison operators
347 can be expressed either through C-like symbols, or through English-like
354 ge, >= Greater than or Equal to
355 le, <= Less than or Equal to
357 Furthermore, each protocol field is typed. The types are:
359 Unsigned integer (either 8-bit, 16-bit, 24-bit, or 32-bit)
360 Signed integer (either 8-bit, 16-bit, 24-bit, or 32-bit)
362 Ethernet address (6 bytes)
363 Byte string (n-number of bytes)
368 Double-precision floating point number
370 An integer may be expressed in decimal, octal, or hexadecimal notation.
371 The following three read filters are equivalent:
377 Boolean values are either true or false. In a read filter expression
378 testing the value of a Boolean field, "true" is expressed as 1 or any
379 other non-zero value, and "false" is expressed as zero. For example, a
380 token-ring packet's source route field is boolean. To find any
381 source-routed packets, a read filter would be:
385 Non source-routed packets can be found with:
389 Ethernet addresses, as well as a string of bytes, are represented in hex
390 digits. The hex digits may be separated by colons, periods, or hyphens:
392 fddi.dst eq ff:ff:ff:ff:ff:ff
393 ipx.srcnode == 0.0.0.0.0.1
394 eth.src == aa-aa-aa-aa-aa-aa
396 If a string of bytes contains only one byte, then it is represented as
397 an unsigned integer. That is, if you are testing for hex value 'ff' in
398 a one-byte byte-string, you must compare it agains '0xff' and not 'ff'.
400 IPv4 addresses can be represented in either dotted decimal notation, or
401 by using the hostname:
403 ip.dst eq www.mit.edu
404 ip.src == 192.168.1.1
406 IPv4 addresses can be compared with the same logical relations as numbers:
407 eq, ne, gt, ge, lt, and le. The IPv4 address is stored in host order,
408 so you do not have to worry about how the endianness of an IPv4 address
409 when using it in a read filter.
411 Classless InterDomain Routing (CIDR) notation can be used to test if an
412 IPv4 address is in a certain subnet. For example, this display filter
413 will find all packets in the 129.111 Class-B network:
415 ip.addr == 129.111.0.0/16
417 Remember, the number after the slash represents the number of bits used
418 to represent the network. CIDR notation can also be used with
419 hostnames, in this example of finding IP addresses on the same Class C
424 The CIDR notation can only be used on IP addresses or hostnames, not in
425 variable names. So, a display filter like "ip.src/24 == ip.dst/24" is
428 IPX networks are represented by unsigned 32-bit integers. Most likely
429 you will be using hexadecimal when testing for IPX network values:
431 ipx.srcnet == 0xc0a82c00
433 A slice operator also exists. You can check the substring
434 (byte-string) of any protocol or field. For example, you can filter on
435 the vendor portion of an ethernet address (the first three bytes) like
438 eth.src[0:3] == 00:00:83
440 If the length of your byte-slice is only one byte, then it is still
441 represented in hex, but without the preceding "0x":
445 You can use the slice operator on a protocol name, too. And
446 remember, the "frame" protocol encompasses the entire packet, allowing
447 you to look at the nth byte of a packet regardless of its frame type
448 (Ethernet, token-ring, etc.).
450 token[0:5] ne 0.0.0.1.1
454 The following syntax governs slices:
456 [i:j] i = start_offset, j = length
457 [i-j] i = start_offet, j = end_offset, inclusive.
458 [i] i = start_offset, length = 1
459 [:j] start_offset = 0, length = j
460 [i:] start_offset = i, end_offset = end_of_field
462 Offsets and lengths can be negative, in which case they indicate the
463 offset from the B<end> of the field. Here's how to check the last 4
466 frame[-4:4] == 0.1.2.3
470 frame[-4:] == 0.1.2.3
472 You can create complex concatenations of slices using the comma operator:
474 field[1,3-5,9:] == 01:03:04:05:09:0a:0b
476 All the above tests can be combined together with logical expressions.
477 These too are expressable in C-like syntax or with English-like
484 Expressions can be grouped by parentheses as well. The following are
485 all valid read filter expression:
487 tcp.port == 80 and ip.src == 192.168.2.1
489 (ipx.srcnet == 0xbad && ipx.srnode == 0.0.0.0.0.1) || ip
490 tr.dst[0:3] == 0.6.29 xor tr.src[0:3] == 0.6.29
492 A special caveat must be given regarding fields that occur more than
493 once per packet. "ip.addr" occurs twice per IP packet, once for the
494 source address, and once for the destination address. Likewise,
495 tr.rif.ring fields can occur more than once per packet. The following
496 two expressions are not equivalent:
498 ip.addr ne 192.168.4.1
499 not ip.addr eq 192.168.4.1
501 The first filter says "show me IP packets where an ip.addr exists that
502 does not equal 192.168.4.1". That is, as long as one ip.addr in the
503 packet does not equal 192.168.44.1, the packet passes the read
504 filter. The second filter "don't show me any packets that have at least
505 one ip.addr field equal to 192.168.4.1". If one ip.addr is 192.168.4.1,
506 the packet does not pass. If B<neither> ip.addr fields is 192.168.4.1,
507 then the packet passes.
509 It is easy to think of the 'ne' and 'eq' operators as having an implict
510 "exists" modifier when dealing with multiply-recurring fields. "ip.addr
511 ne 192.168.4.1" can be thought of as "there exists an ip.addr that does
512 not equal 192.168.4.1".
514 Be careful with multiply-recurring fields; they can be confusing.
516 Care must also be taken when using the read filter to remove noise
517 from the packet trace. If you want to e.g. filter out all IP multicast
518 packets to address 224.1.2.3, then using:
522 may be too restrictive. Filtering with "ip.dst" selects only those
523 B<IP> packets that satisfy the rule. Any other packets, including all
524 non-IP packets, will not printed. For printing also the non-IP
525 packets, you can use one of the following two expressions:
527 not ip or ip.dst ne 224.1.2.3
528 not ip.addr eq 224.1.2.3
530 The first filter uses "not ip" to include all non-IP packets and then
531 lets "ip.dst ne 224.1.2.3" to filter out the unwanted IP packets. The
532 second filter has already been explained above where filtering with
533 multiply occuring fields was discussed.
535 The following is a table of protocol and protocol fields that are
536 filterable in B<Tethereal>. The abbreviation of the protocol or field is
537 given. This abbreviation is what you use in the read filter. The
538 type of the field is also given.
540 =insert_dfilter_table
544 The F<ethereal.conf> file, which is installed in the F<etc> directory
545 under the main installation directory (for example, F</usr/local/etc>)
546 on UNIX-compatible systems, and in the main installation directory (for
547 example, F<C:\Program Files\Ethereal>) on Windows systems, and the
548 personal preferences file, which is F<$HOME/.ethereal/preferences> on
549 UNIX-compatible systems and F<%APPDATA%\Ethereal\preferences> (or, if
550 %APPDATA% isn't defined,
551 F<%USERPROFILE%\Application Data\Ethereal\preferences>) on
552 Windows systems, contain system-wide and personal preference settings,
553 respectively. The file contains preference settings of the form
554 I<prefname>B<:>I<value>, one per line, where I<prefname> is the name of
555 the preference (which is the same name that would appear in the
556 preference file), and I<value> is the value to which it should be set;
557 white space is allowed between B<:> and I<value>. A preference setting
558 can be continued on subsequent lines by indenting the continuation lines
559 with white space. A B<#> character starts a comment that runs to the
562 The system-wide preference file is read first, if it exists, overriding
563 B<Tethereal>'s default values; the personal preferences file is then
564 read, if it exists, overriding default values and values read from the
565 system-wide preference file.
567 The F<ethers> file, which is found in the F</etc> directory on
568 UNIX-compatible systems, and in the main installation directory (for
569 example, F<C:\Program Files\Ethereal>) on Windows systems, is consulted
570 to correlate 6-byte hardware addresses to names. If an address is not
571 found in the F<ethers> file, the F<$HOME/.ethereal/ethers> file on
572 UNIX-compatible systems, and the F<%APPDATA%\Ethereal\ethers> file (or, if
573 %APPDATA% isn't defined, the
574 F<%USERPROFILE%\Application Data\Ethereal\ethers> file) on Windows
575 systems is consulted next. Each line contains one hardware
576 address and name, separated by whitespace. The digits of the hardware
577 address are separated by either a colon (:), a dash (-), or a period
578 (.). The following three lines are valid lines of an ethers file:
580 ff:ff:ff:ff:ff:ff Broadcast
581 c0-00-ff-ff-ff-ff TR_broadcast
582 00.00.00.00.00.00 Zero_broadcast
584 The F<manuf> file, which is installed in the F<etc> directory under the
585 main installation directory (for example, F</usr/local/etc>) on
586 UNIX-compatible systems, and in the main installation directory (for
587 example, F<C:\Program Files\Ethereal>) on Windows systems, matches the
588 3-byte vendor portion of a 6-byte hardware address with the
589 manufacturer's name; it can also contain well-known MAC addresses and
590 address ranges specified with a netmask. The format of the file is the
591 same as the F<ethers> file, except that entries of the form
595 can be provided, with the 3-byte OUI and the name for a vendor, and
598 00-00-0C-07-AC/40 All-HSRP-routers
600 can be specified, with a MAC address and a mask indicating how many bits
601 of the address must match. Trailing zero bytes can be omitted from
602 address ranges. That entry, for example, will match addresses from
603 00-00-0C-07-AC-00 through 00-00-0C-07-AC-FF. The mask need not be a
606 The F<ipxnets> file, which is found in the F</etc> directory on
607 UNIX-compatible systems, and in the main installation directory (for
608 example, F<C:\Program Files\Ethereal>) on Windows systems, correlates
609 4-byte IPX network numbers to names. If a network number is not found
610 in the F<ipxnets> file, the F<$HOME/.ethereal/ipxnets> file on
611 UNIX-compatible systems, and the F<%APPDATA%\Ethereal\ipxnets> file (or,
612 if %APPDATA% isn't defined, the
613 F<%USERPROFILE%\Application Data\Ethereal\ipxnets> file)
614 on Windows systems, is consulted next. The format is the same as the
615 F<ethers> file, except that each address if four bytes instead of six.
616 Additionally, the address can be represented a single hexadecimal
617 number, as is more common in the IPX world, rather than four hex octets.
618 For example, these four lines are valid lines of an ipxnets file.
622 00:00:BE:EF IT_Server1
627 I<ethereal(1)>, I<editcap(1)>, I<tcpdump(8)>, I<pcap(3)>
631 B<Tethereal> is part of the B<Ethereal> distribution. The latest version
632 of B<Ethereal> can be found at B<http://www.ethereal.com>.
636 B<Tethereal> uses the same packet dissection code that B<Ethereal> does,
637 as well as using many other modules from B<Ethereal>; see the list of
638 authors in the B<Ethereal> man page for a list of authors of that code.