tcpflow - TCP flow recorder

tcpflow [c -cCehpsvc ] [c -b max_bytesc ] [c -d debug_levelc ] [c -f max_fdsc ] [c -i ifacec ] [c -r filec ] [c expressionc ]

tcpflow is a program that captures data transmitted as part of TCP connections (flows), and stores the data in a way that is convenient for protocol analysis or debugging. A program like tcpdump(4) shows a summary of packets seen on the wire, but usually doesn't store the data that's actually being transmitted. In contrast, tcpflow reconstructs the actual data streams and stores each flow in a separate file for later analysis. tcpflow understands TCP sequence numbers and will correctly reconstruct data streams regardless of retransmissions or out-of-order delivery.

tcpflow stores all captured data in files that have names of the form

192.168.101.102.02345-010.011.012.013.45103

-b

Max bytes per flow. Capture no more than max_bytes bytes per flow. Any data captured for a flow beyond max_bytes from the first byte captured will be discarded. The default is to store an unlimited number of bytes per flow.

-c

Console print. Print the contents of packets to stdout as they are received, without storing any captured data to files (implies

-C

Console print without the packet source and destination details being printed. Print the contents of packets to stdout as they are received, without storing any captured data to files (implies -e When outputting to the console each flow will be output in alternating colours. -s ).

-d

Debug level. Set the level of debugging messages printed to stderr to debug_level. Higher numbers produce more messages. -d 0 causes completely silent operation. -d 1 , the default, produces minimal status messages. -d 10 produces verbose output equivalent to -v . Numbers higher than 10 can produce a large amount of debugging information useful only to developers.

-f

Max file descriptors used. Limit the number of file descriptors used by tcpflow to max_fds. Higher numbers use more system resources, but usually perform better. If the underlying operating system supports the setrlimit() system call, the OS will be asked to enforce the requested limit. The default is for tcpflow to use the maximum number of file descriptors allowed by the OS. The -v option will report how many file descriptors tcpflow is using.

-h

Help. Print usage information and exit.

-i

Interface name. Capture packets from the network interface named iface. If no interface is specified with -i , a reasonable default will be used by libpcap automatically.

-p

No promiscuous mode. Normally, tcpflow attempts to put the network interface into promiscuous mode before capturing packets. The -p option tells tcpflow not to put the interface into promiscuous mode. Note that it might already be in promiscuous mode for some other reason.

-r

Read from file. Read packets from file, which was created using the -w option of tcpdump(1). Standard input is used if file is ``-''. Note that for this option to be useful, tcpdump's -s option should be used to set the snaplen to the MTU of the interface (e.g., 1500) while capturing packets.

-s

Strip non-printables. Convert all non-printable characters to the "." character before printing packets to the console or storing them to a file.

-v

Verbose operation. Verbosely describe tcpflow's operation. Equivalent to -d 10 .

The expression specified on the command-line specifies which packets should be captured. Because tcpflow uses the the libpcap library, tcpflow has the same powerful filtering language available as programs such as tcpdump(1).

The following part of the man page is excerpted from the tcpdump man page.

expression selects which packets will be dumped. If no expression is given, all packets on the net will be dumped. Otherwise, only packets for which expression is `true' will be dumped.

The expression consists of one or more primitives. Primitives usually consist of an id (name or number) preceded by one or more qualifiers. There are three different kinds of qualifier:

type

qualifiers say what kind of thing the id name or number refers to. Possible types are host, net and port. E.g., `host foo', `net 128.3', `port 20'. If there is no type qualifier, host is assumed.

dir

qualifiers specify a particular transfer direction to and/or from id. Possible directions are src, dst, src or dst and src and dst. E.g., `src foo', `dst net 128.3', `src or dst port ftp-data'. If there is no dir qualifier, src or dst is assumed. For `null' link layers (i.e. point to point protocols such as slip) the inbound and outbound qualifiers can be used to specify a desired direction.

proto

qualifiers restrict the match to a particular protocol. Possible protos are: ether, fddi, ip, arp, rarp, decnet, lat, sca, moprc, mopdl, tcp and udp. E.g., `ether src foo', `arp net 128.3', `tcp port 21'. If there is no proto qualifier, all protocols consistent with the type are assumed. E.g., `src foo' means `(ip or arp or rarp) src foo' (except the latter is not legal syntax), `net bar' means `(ip or arp or rarp) net bar' and `port 53' means `(tcp or udp) port 53'.

[`fddi' is actually an alias for `ether'; the parser treats them identically as meaning ``the data link level used on the specified network interface.'' FDDI headers contain Ethernet-like source and destination addresses, and often contain Ethernet-like packet types, so you can filter on these FDDI fields just as with the analogous Ethernet fields. FDDI headers also contain other fields, but you cannot name them explicitly in a filter expression.]

In addition to the above, there are some special `primitive' keywords that don't follow the pattern: gateway, broadcast, less, greater and arithmetic expressions. All of these are described below.

More complex filter expressions are built up by using the words and, or and not to combine primitives. E.g., `host foo and not port ftp and not port ftp-data'. To save typing, identical qualifier lists can be omitted. E.g., `tcp dst port ftp or ftp-data or domain' is exactly the same as `tcp dst port ftp or tcp dst port ftp-data or tcp dst port domain'.

Allowable primitives are:

dst host host

True if the IP destination field of the packet is host, which may be either an address or a name.

src host host

True if the IP source field of the packet is host.

"host True if either the IP source or destination of the packet is host. Any of the above host expressions can be prepended with the keywords, ip, arp, or rarp as in:

ip host host

which is equivalent to:

ether proto \ip and host host

If host is a name with multiple IP addresses, each address will be checked for a match.

"ether True if the ethernet destination address is ehost. Ehost may be either a name from /etc/ethers or a number (see ethers(3N) for numeric format).

"ether True if the ethernet source address is ehost.

"ether True if either the ethernet source or destination address is ehost.

"gateway

True if the packet used host as a gateway. I.e., the ethernet source or destination address was host but neither the IP source nor the IP destination was host. Host must be a name and must be found in both /etc/hosts and /etc/ethers. (An equivalent expression is

ether host ehost and not host host

which can be used with either names or numbers for host / ehost.)

dst net net

True if the IP destination address of the packet has a network number of net. Net may be either a name from /etc/networks or a network number (see networks(5) for details).

src net net

True if the IP source address of the packet has a network number of net.

net net

True if either the IP source or destination address of the packet has a network number of net.

net net mask mask

True if the IP address matches net with the specific netmask. May be qualified with src or dst.

net net/len

True if the IP address matches net a netmask len bits wide. May be qualified with src or dst.

dst port port

True if the packet is ip/tcp or ip/udp and has a destination port value of port. The port can be a number or a name used in /etc/services (see tcp(4P) and udp(4P)). If a name is used, both the port number and protocol are checked. If a number or ambiguous name is used, only the port number is checked (e.g., dst port 513 will print both tcp/login traffic and udp/who traffic, and port domain will print both tcp/domain and udp/domain traffic).

src port port

True if the packet has a source port value of port.

port port

True if either the source or destination port of the packet is port. Any of the above port expressions can be prepended with the keywords, tcp or udp, as in:

tcp src port port

which matches only tcp packets whose source port is port.

less length

True if the packet has a length less than or equal to length. This is equivalent to:

len <= length.

greater length

True if the packet has a length greater than or equal to length. This is equivalent to:

len >= length.

ip proto protocol

True if the packet is an ip packet (see ip(4P)) of protocol type protocol. Protocol can be a number or one of the names icmp, igrp, udp, nd, or tcp. Note that the identifiers tcp, udp, and icmp are also keywords and must be escaped via backslash (\), which is \\ in the C-shell.

ether broadcast

True if the packet is an ethernet broadcast packet. The ether keyword is optional.

ip broadcast

True if the packet is an IP broadcast packet. It checks for both the all-zeroes and all-ones broadcast conventions, and looks up the local subnet mask.

ether multicast

True if the packet is an ethernet multicast packet. The ether keyword is optional. This is shorthand for `ether[0] & 1 != 0'.

ip multicast

True if the packet is an IP multicast packet.

ether proto protocol

True if the packet is of ether type protocol. Protocol can be a number or a name like ip, arp, or rarp. Note these identifiers are also keywords and must be escaped via backslash (\). [In the case of FDDI (e.g., `fddi protocol arp'), the protocol identification comes from the 802.2 Logical Link Control (LLC) header, which is usually layered on top of the FDDI header. Tcpdump assumes, when filtering on the protocol identifier, that all FDDI packets include an LLC header, and that the LLC header is in so-called SNAP format.]

decnet src host

True if the DECNET source address is host, which may be an address of the form ``10.123'', or a DECNET host name. [DECNET host name support is only available on Ultrix systems that are configured to run DECNET.]

decnet dst host

True if the DECNET destination address is host.

decnet host host

True if either the DECNET source or destination address is host.

ip, arp, rarp, decnet

Abbreviations for:

ether proto p

where p is one of the above protocols.

lat, moprc, mopdl

Abbreviations for:

ether proto p

where p is one of the above protocols. Note that tcpdump does not currently know how to parse these protocols.

tcp, udp, icmp

Abbreviations for:

ip proto p

where p is one of the above protocols.

expr relop expr

True if the relation holds, where relop is one of >, <, >=, <=, =, !=, and expr is an arithmetic expression composed of integer constants (expressed in standard C syntax), the normal binary operators [+, -, *, /, &, |], a length operator, and special packet data accessors. To access data inside the packet, use the following syntax:

proto [ expr : size ]

Proto is one of ether, fddi, ip, arp, rarp, tcp, udp, or icmp, and indicates the protocol layer for the index operation. The byte offset, relative to the indicated protocol layer, is given by expr. Size is optional and indicates the number of bytes in the field of interest; it can be either one, two, or four, and defaults to one. The length operator, indicated by the keyword len, gives the length of the packet.

For example, `ether[0] & 1 != 0' catches all multicast traffic. The expression `ip[0] & 0xf != 5' catches all IP packets with options. The expression `ip[6:2] & 0x1fff = 0' catches only unfragmented datagrams and frag zero of fragmented datagrams. This check is implicitly applied to the tcp and udp index operations. For instance, tcp[0] always means the first byte of the TCP header, and never means the first byte of an intervening fragment.

Primitives may be combined using:

A parenthesized group of primitives and operators (parentheses are special to the Shell and must be escaped).

Negation (`!' or `not').

Concatenation (`&&' or `and').

Alternation (`||' or `or').

Negation has highest precedence. Alternation and concatenation have equal precedence and associate left to right. Note that explicit and tokens, not juxtaposition, are now required for concatenation.

If an identifier is given without a keyword, the most recent keyword is assumed. For example,

not host vs and ace

not host vs and host ace

not ( host vs or ace )

Expression arguments can be passed to tcpdump as either a single argument or as multiple arguments, whichever is more convenient. Generally, if the expression contains Shell metacharacters, it is easier to pass it as a single, quoted argument. Multiple arguments are concatenated with spaces before being parsed.

The following part of the man page is excerpted from the tcpdump man page.

To record all packets arriving at or departing from sundown:

tcpflow host sundown

To record traffic between helios and either hot or ace:

tcpflow host helios and \( hot or ace \)

To record traffic between ace and any host except helios:

tcpflow host ace and not helios

To record all traffic between local hosts and hosts at Berkeley:

tcpflow net ucb-ether

To record all ftp traffic through internet gateway snup: (note that the expression is quoted to prevent the shell from (mis-)interpreting the parentheses):

tcpflow 'gateway snup and (port ftp or ftp-data)'

Please send bug reports to jelson@circlemud.org.

tcpflow currently does not understand IP fragments. Flows containing IP fragments will not be recorded correctly.

tcpflow never frees state associated with flows that it records, so will grow large if used to capture a very large number of flows (e.g., on the order of 100,000 flows or more).

There appears to be a bug in the way that Linux delivers packets to libpcap when using the loopback interface ("localhost"). When listening to the Linux loopback interface, selective packet filtering is not possible; all TCP flows on the localhost interface will be recorded.

Jeremy Elson <jelson@circlemud.org>

The current version of this software is available at http://www.circlemud.org/~jelson/software/tcpflow

tcpdump(1), nit(4), bpf(4), pcap(3)