• 1 Introduction
• 2 Common options
• 3 dnsdomainname: Show DNS domain name
• 4 hostname: Show or set system host name.
• 5 ifconfig: Configure network interfaces
• 6 logger: Send messages to system log
• 7 ping: Packets to network hosts
• 8 ping6: Packets to IPv6 network hosts
• 9 traceroute: Trace the route to a host
• 10 whois: User interface to WHOIS data bases.
• 11 ftp: FTP client
• 12 rcp: Copy files between machines
• 13 rexec: a remote execution program
• 14 rlogin: Remote login
• 15 rsh: Remote shell
• 16 talk: a communication program
• 17 telnet: User interface to TELNET
• 18 tftp: TFTP client
• 19 inetd: Internet super-server
• 20 syslogd: system service logging faclity
• 21 ftpd: FTP daemon
• 22 rexecd: server for rexec
• 23 rlogind: Remote login server
• 24 rshd: Remote shell server
• 25 talkd: a server for communication between users
• 26 telnetd: Telnet server
• 27 tftpd: TFTP server
• 28 uucpd: Unix to Unix Copy relay daemon.
• Appendix A GNU Free Documentation License
• Index

• 1 Introduction
• 2 Common options
  • 2.1 Exit status
• 3 dnsdomainname: Show DNS domain name
• 4 hostname: Show or set system host name.
  • 4.1 Command line options
• 5 ifconfig: Configure network interfaces
  • 5.1 Command line options
  • 5.2 Formatted status output
  • 5.3 Legacy syntax
• 6 logger: Send messages to system log
  • 6.1 Command line options
  • 6.2 Examples
• 7 ping: Packets to network hosts
  • 7.1 Command line options
  • 7.2 Using ping for network fault isolation
  • 7.3 Duplicate and damaged packets
  • 7.4 Trying different data patterns
  • 7.5 TTL details
  • 7.6 Further observations
• 8 ping6: Packets to IPv6 network hosts
  • 8.1 Command line options
• 9 traceroute: Trace the route to a host
  • 9.1 Command line options
  • 9.2 Diagnostic tokens
• 10 whois: User interface to WHOIS data bases.
  • 10.1 Command line options
  • 10.2 Environment variables
• 11 ftp: FTP client
  • 11.1 Command line options
  • 11.2 Commands interpreted by ftp
  • 11.3 Environment variables in use
  • 11.4 Aborting a file transfer
  • 11.5 File naming conventions
  • 11.6 File transfer parameters
  • 11.7 The .netrc file
• 12 rcp: Copy files between machines
  • 12.1 Command line options
• 13 rexec: a remote execution program
  • 13.1 Command line options
• 14 rlogin: Remote login
  • 14.1 Command line options
  • 14.2 Escape characters and flow control
  • 14.3 Kerberos Authentication
• 15 rsh: Remote shell
  • 15.1 Command line options
  • 15.2 Note on stream redirections
• 16 talk: a communication program
  • 16.1 Invoking
• 17 telnet: User interface to TELNET
  • 17.1 Command line options
• 18 tftp: TFTP client
  • 18.1 Commands
• 19 inetd: Internet super-server
  • 19.1 Invocation
  • 19.2 Configuration file
  • 19.3 Built-in services
  • 19.4 TCPMUX
  • 19.5 Inetd Environment
  • 19.6 Error Messages
• 20 syslogd: system service logging faclity
  • 20.1 Configuration file
• 21 ftpd: FTP daemon
  • 21.1 Standards
  • 21.2 Authentication
  • 21.3 Configuration files
  • 21.4 File format of ftpusers and ftpchroot.
• 22 rexecd: server for rexec
  • 22.1 Invoking
  • 22.2 Diagnostics
• 23 rlogind: Remote login server
  • 23.1 Invoking
  • 23.2 Kerberos specific details
  • 23.3 Protocol details
  • 23.4 Diagnostics
• 24 rshd: Remote shell server
  • 24.1 Invoking
  • 24.2 Diagnostics
• 25 talkd: a server for communication between users
  • 25.1 Invoking
  • 25.2 Modus operandi
  • 25.3 Access control in talkd
• 26 telnetd: Telnet server
  • 26.1 Crafting an execution string.
• 27 tftpd: TFTP server
  • 27.1 Directory prefixes
  • 27.2 Use cases
• 28 uucpd: Unix to Unix Copy relay daemon.
  • 28.1 Options
  • 28.2 Authentication steps.
• Appendix A GNU Free Documentation License
• Index

Up: Common options   [Contents][Index]

2.1 Exit status ¶

Nearly every command invocation yields an integral exit status that can be used to change how other commands work. For the vast majority of commands, an exit status of zero indicates success. Failure is indicated by a nonzero value — typically ‘1’, though it may differ on unusual platforms, as POSIX requires only that it be nonzero.

dnsdomainname [option...]

hostname [option...]
hostname name

4.1 Command line options ¶

-a ¶

--aliases

Get alias names.

-d ¶

--domain

Get DNS domain name.

-f ¶

--fqdn

--long

Get DNS host name or Fully Qualified Domain Name.

-F file ¶

--file=file

Set host name or NIS domain name from FILE.

-i ¶

--ip-addresses

Get addresses for the host name.

-s ¶

--short

Get short host name.

-y ¶

--yp

--nis

Get NIS/YP domain name.

ifconfig iface [arg...]
ifconfig -i iface [option...] [-i iface2 [option...]]

5.1 Command line options ¶

-a ¶

--all

Display all available interfaces, including those that not are marked as ‘up’, i.e., also the inactive interfaces.

-A addr ¶

--address=addr

Set address of selected interface to addr.

-b addr ¶

-B addr

--brdaddr=addr

--broadcast=addr

Set broadcast address of selected interface to addr.

-d addr ¶

-p addr

--dstaddr=addr

--peer=addr

Set destination (peer) address of selected interface.

--down ¶

Deactivate the selected interface.

-F list ¶

--flags=list

Change those interface flags mentioned in list. The argument is a comma separated list of one or more flag names to be set, or in case the name is prepended with ‘no’, the corresponding flag is cleared. The output of ifconfig with the option --help contains a list of available flag names.

--format=format ¶

Select output format; the value ‘help’ prints a list of all available formats.

-i name ¶

--interface=name

Select the named interface for any following action.

-l ¶

--list

List, with name only, all available interfaces, or only those selected should at least one option -i have specified.

-m mask ¶

--netmask=mask

Set netmask of selected interface to mask.

--metric=n ¶

Set the metric of selected interface to the number n.

-M n ¶

--mtu=n

Set MTU of selected interface to the number n.

-s ¶

--short

Use short output format. This is identical to specifying ‘--format=netstat’.

--up ¶

Activate the selected interface.

-v ¶

--verbose

Print informational messages when configuring an interface.

Observe that the use of program options is the only manner in which ifconfig is able to handle multiple interfaces in one invocation. Once a particular interface has been selected using -i, it is affected by any following option until replaced by another interface selector. This is also the main cause, that ifconfig is unable to treat options independently of their order, as is mostly the case in other GNU software.

5.2 Formatted status output ¶

The status of one or more interfaces can be presented in a number of different formats. A list of them is printed by the option --format=help. In the following table the valid formats are given, each is used in the form --format=name.

check

check-existence

?

Place holders for the ability to check whether the interfaces selected by one or more options -i are determining existing interfaces in the running system. No output in case of success, an error message in case of a failure.

gnu

default

Standard GNU output format.

gnu-one-entry

Like the previous format, but with intermediary newlines removed.

help

Display a list of valid formats, together with a short description for each choice.

net-tools

Imitation of presentation used by the implementation in ‘net-tools’. Default format for GNU/Linux.

netstat

Terse output with statistics, similar to that of netstat -i.

osf

Format variant of ‘unix’ preferred by OSF’s implementation.

unix

Traditional UNIX type format. Default for BSD, HPUX and Solaris.

5.3 Legacy syntax ¶

The traditional mode of invoking ifconfig is via a parsed command line, without all use of program switches and options, relying fully on argument parsing. This mode of use is supported also in the present implementation, but keep in mind that only one interface can be manipulated using this legacy syntax.

ifconfig NAME [ADDR [DSTADDR]] [broadcast BRDADDR] [netmask MASK]
         [metric N] [mtu N] [up|down]

As is conventional, only the primary address and possibly the peer destination address are stated as bare arguments, without a specifying keyword. Some slight variation on this syntax will depend on the target system for which the program is being built, as not all platforms support identical abilities. The best information is found via the usage massage ‘ifconfig --usage’.

logger [option...] [message]

6.1 Command line options ¶

-4 ¶

--ipv4

Use IPv4 as transport when logging to a host. The default behaviour is to use whatever IP version that matches the host.

-6 ¶

--ipv6

Use IPv6 as transport when logging to a host. The option is present also on systems without support for IPv6, but will then issue a warning and then fall back to IPv4 when delivering the message.

Both options are most influential when the target host is named using a symbolic name, but numerical addresses for host or source must also match if either of --ipv4 or --ipv6 is stated.

-f file ¶

--file=file

Log the content of the specified file. If file is ‘-’ then standard input is assumed.

-h host ¶

--host=host

Send messages to the given host or socket. The host argument can be either a local UNIX socket name (containing a slash ‘/’), or be of the form

where host is the remote host name or IP address, and the optional port is a decimal port number or symbolic service name from /etc/services. If port is not specified, the port number corresponding to the ‘syslog’ service is used. If a numerical IPv6 address is given without a port specification, then the address must be enclosed within brackets (like [::1]).

-i[pid] ¶

--id=[pid]

Add process ID to each message. If pid is not supplied, use the process ID of the logger process with each line. Notice, that pid is an optional argument. When supplied to the -i option, it must follow the ‘i’ letter immediately, without any separating whitespace. When supplied to the --id form, it must be separated from it by exactly one equals sign.

-p priority ¶

--priority=priority

Enter the message with the specified priority. The priority may be specified numerically or as a ‘facility.level’ pair. For example, -p local3.info logs the message at the informational level in the ‘local3’ facility. The default is ‘user.notice’.

The actual list of supported facilities and levels is system specific.

-s ¶

--stderr

Log the message to standard error, as well as to the system log.

-S addr ¶

--source=addr

Supply the source IP address for INET connections. This option is useful in conjunction with --host (see above). The kind of address specified here (IPv4 or IPv6) will propagate to influence the resolution of the host address, if it is a symbolic name.

-t tag ¶

--tag=tag

Mark every line in the log with the specified tag.

-u socket ¶

--unix=socket

Send messages to the given local UNIX socket. The socket argument can be either an absolute path (starting with a slash ‘/’), or a relative path understood relative to the current working directory.

The options are followed by the message which should be written to the log. If not specified, and the -f flag is not provided, standard input is logged.

6.2 Examples ¶

The following examples illustrate the usage of the logger command:

1. Log the message ‘System rebooted’ to the local syslog. Use default facility and priority:
2. Run command and send its error output to the channel ‘local0.err’. Mark each message with tag ‘cmd’:

   command 2>&1 | logger -p local0.err -t cmd
   

3. Log each line from file warnings to channel ‘daemon.warn’ on host ‘logger.runasimi.org’, using the source IP ‘10.10.10.1’:

   logger -p daemon.warn -h logger.runasimi.org -S 10.10.10.1  \
          --file warnings
   

Next: Using ping for network fault isolation, Up: ping: Packets to network hosts   [Contents][Index]

7.1 Command line options ¶

Selection of packet type is handled by these first options:

--address ¶

Send ICMP_ADDRESS packets, thus requesting the address netmask in use by the targeted host.

--echo ¶

Send ICMP_ECHO requests. This is the default action.

--mask ¶

Identical to --address.

--timestamp ¶

Send ICMP_TIMESTAMP packets, thereby requesting a timed response from the targeted host.

In successful cases three time values are returned. All are expected to state the number of milliseconds since midnight UTC. The first of these, ‘icmp_otime’, contains the original time of sending the request. Then comes ‘icmp_rtime’, the time of reception by the target, and finally, ‘icmp_ttime’, the time of transmitting an answer back to the originator.

-t type ¶

--type=type

Send type packets. Accepted values are ‘address’, ‘echo’, ‘mask’, and ‘timestamp’.

The following options are available for all packet types:

-c n ¶

--count=n

Stop after sending and receiving answers to a total of n packets.

-d ¶

--debug

Set the SO_DEBUG option on the socket being used.

-i n ¶

--interval=n

Wait n seconds until sending next packet. The default is to wait for one second between packets. This option is incompatible with the option -f.

-n ¶

--numeric

Numeric output only. No attempt will be made to resolve symbolic names for host addresses.

-r ¶

--ignore-routing

Bypass the normal routing tables and send directly to a host on an attached network. If the host is not on a directly attached network, an error is returned. This option can be used to ping a local host through an interface that has no route through it (e.g., after the interface was dropped by routed).

-T num ¶

--tos=num

Set type-of-service, TOS field, to num on transmitted packets.

--ttl=n ¶

Set the specified number n as value of time-to-live when transmitting packets. Acceptable values are 1 to 255, inclusive.

-v ¶

--verbose

Produce more verbose output, giving more statistics.

-w n ¶

--timeout=n

Stop after n seconds.

-W n ¶

--linger=n

Maximum number of seconds n to wait for a response.

Finally, these last options are relevant only for sending echo requests, allowing many variations in order to detect various peculiarities of the targeted host, or the intermediary routers for that matter.

-f ¶

--flood

Flood ping. Outputs packets as fast as they come back or one hundred times per second, whichever is more. For every ECHO_REQUEST packet sent, a period ‘.’ is printed, while for every ECHO_REPLY received in reply, a backspace is printed. This provides a rapid display of how many packets are being dropped. Only the super-user may use this option. This can be very hard on a network and should be used with caution.

--ip-timestamp=flag ¶

Include IP option Timestamp in transmitted packets. The value flag is either ‘tsonly’, which only records up to nine time stamps, or ‘tsaddr’, which records IP addresses as well as time stamps, but for at most four hosts.

-l n ¶

--preload=n

If n is specified, ping sends that many packets as fast as possible before falling into its normal mode of operation.

-p pat ¶

--pattern=pat

You may specify up to 16 pad bytes to fill out the packet you send. This is useful for diagnosing data-dependent problems in a network. For example, -p ff will cause the sent packet to be filled with all ones.

-q ¶

--quiet

Do not print timing for each transmitted packet.

-R ¶

--route

Record route. Includes the RECORD_ROUTE field in the ECHO_REQUEST packet and displays the route buffer on returned packets. Note that the IP header is only large enough for nine such routes. Many hosts ignore or discard this option.

-s n ¶

--size=n

Specifies the number of data bytes to be sent. The default is 56, which translates into 64 ICMP data bytes, taking the 8 bytes of ICMP header data into account.

Next: Duplicate and damaged packets, Previous: Command line options, Up: ping: Packets to network hosts   [Contents][Index]

7.2 Using ping for network fault isolation ¶

When using ping for fault isolation, it should first be run on the local host, to verify that the local network interface is up and running. Then, hosts and gateways further and further away should be pinged. Round-trip times and packet loss statistics are computed. If duplicate packets are received, they are not included in the packet loss calculation, although the round trip time of these packets is used in calculating the minimum/average/maximum round-trip time numbers. When the specified number of packets have been sent (and received) or if the program is terminated with a ‘SIGINT’, a brief summary is displayed.

This program is intended for use in network testing, measurement and management. Because of the load it can impose on the network, it is unwise to use ping during normal operations or from automated scripts.

Next: Trying different data patterns, Previous: Using ping for network fault isolation, Up: ping: Packets to network hosts   [Contents][Index]

7.3 Duplicate and damaged packets ¶

Ping will report duplicate and damaged packets. Duplicate packets should never occur, and seem to be caused by inappropriate link-level retransmissions. Duplicates may occur in many situations and are rarely (if ever) a good sign, although the presence of low levels of duplicates may not always be cause for alarm.

Damaged packets are obviously serious cause for alarm and often indicate broken hardware somewhere in the ping packet’s path (in the network or in the hosts).

Next: TTL details, Previous: Duplicate and damaged packets, Up: ping: Packets to network hosts   [Contents][Index]

7.4 Trying different data patterns ¶

The (inter)network layer should never treat packets differently depending on the data contained in the data portion. Unfortunately, data-dependent problems have been known to sneak into networks and remain undetected for long periods of time. In many cases the particular pattern that will have problems is something that doesn’t have sufficient “transitions”, such as all ones or all zeros, or a pattern right at the edge, such as almost all zeros. It isn’t necessarily enough to specify a data pattern of all zeros (for example) on the command line because the pattern that is of interest is at the data link level, and the relationship between what you type and what the controllers transmit can be complicated.

This means that if you have a data-dependent problem you will probably have to do a lot of testing to find it. If you are lucky, you may manage to find a file that either can’t be sent across your network or that takes much longer to transfer than other similar length files. You can then examine this file for repeated patterns that you can test using the -p option of ping.

Next: Further observations, Previous: Trying different data patterns, Up: ping: Packets to network hosts   [Contents][Index]

7.5 TTL details ¶

The TTL field, Time To Live, of an IP packet represents the maximum number of IP routers that the packet can go through before being discarded. In current practice you can expect each router on the Internet to decrement the TTL field by exactly one.

The TCP/IP specification states that the TTL field of a new TCP packet should be set to 60, but many systems use smaller values (4.3BSD used 30 and 4.2BSD used 15).

The maximum possible value of this field is 255, and most UNIX systems set the TTL field of ICMP (type ECHO_REQUEST) packets to 255. This is why you will find you can ping some hosts, but not reach them with telnet or ftp.

During normal operation, ping prints the TTL value for every packet it receives. When a remote system receives an ICMP packet, it can do one of three things to the TTL field in its response packet:

• Not to change it. This is what Berkeley UNIX systems did before the 4.3BSD-Tahoe release. In this case the TTL value in the received packet will be 255 minus the number of routers in the round-trip path.
• Set it to 255. This is what current Berkeley UNIX systems do. In this case the TTL value in the received packet will be 255 minus the number of routers in the path from the remote system to the pinging host.
• Set it to some other value. Some machines use the same value for ICMP packets that they use for TCP packets, for example either 30 or 60. Others may use completely arbitrary values.

8.1 Command line options ¶

-c n ¶

--count=n

Stop after sending and receiving answers to a total of n packets.

-d ¶

--debug

Set the SO_DEBUG option on the socket being used.

-f ¶

--flood

Flood ping. Outputs packets as fast as they come back, or one hundred times per second, whichever is more. For every ECHO_REQUEST packet sent, a period ‘.’ is printed, while for every ECHO_REPLY received in reply, a backspace is printed.

This provides a rapid display of how many packets are being dropped. Only the super-user may use this option. This mode can be very hard on a network. It should be used with caution!

--hoplimit=n ¶

Limit maximal distance to n. Acceptable values are 1 to 255, inclusive.

-i n ¶

--interval=n

Wait n seconds until sending next packet. The default is to wait for one second between packets. This option is incompatible with the option -f.

-l n ¶

--preload=n

Sends n packets as fast as possible before falling back to the normal mode of operation.

-n ¶

--numeric

Numeric output only. No attempt will be made to resolve symbolic names for host addresses.

-p pattern ¶

--pattern=pattern

Up to 16 hexadecimal pad bytes are given as pattern. These are use for filling out the packets you send. This option is useful for diagnosing data-dependent problems within a network. As an example, -p ff will cause the sent packets to have payloads with every bit set to one.

-q ¶

--quiet

Do not print timing result of each transmitted packet.

-r ¶

--ignore-routing

Bypass the normal routing tables and send directly to a host on an attached network. If the host is not on a directly attached network, an error is returned. This option can be used to ping a local host through an interface, for which there is no assigned route, such as when the interface was dropped by routed.

-s n ¶

--size=n

Specifies the number of data bytes to be sent. The default is 56, which translates into 64 ICMP data bytes, taking the 8 bytes of ICMP header data into account.

-T num ¶

--tos=num

Set the traffic class to num on transmitted packets.

--ttl=n ¶

Synonym for --hoplimit.

-v ¶

--verbose

Produce more verbose output, giving more statistics.

-w n ¶

--timeout=n

Stop after n seconds.

The documentation of ping provides several pieces of information, and discussions, relevant to the use of ping6. Keep in mind, though, that the differing address family causes some discrepancy. See ping: Packets to network hosts.

traceroute [option...] host

9.1 Command line options ¶

-f num ¶

--first-hop=num

Set the initial hop distance to num, instead of the default 1. This immediately allows probing packets to sense routing properties closer to the target host, skipping routers close to the local host. Quicker analysis of problems known to lie at some routing distance is the outcome.

-g gates ¶

--gateways=gates

Set intermediary hosts used in loose source routing. The argument gates is a list of gateways, using space, comma, or semicolon as separators. These hosts must be traversed in the given order before the intended host receives any datagram. At most eight host names or addresses may be specified. Multiple uses of -g produce a concatenated list.

-I ¶

--icmp

Use ICMP ECHO datagrams for probing the remote host.

-m num ¶

--max-hop=num

Set the maximum time-to-live allowed for probing. In other words, stop probing when the hop distance is in excess of num. The default limit is 64.

-M method ¶

--type=method

Use method as carrier packets for traceroute operations. Supported choices are ‘icmp’ and ‘udp’, where ‘udp’ is the default type.

-p port ¶

--port=port

Set destination port of target to port. The default value is 33434.

-q num ¶

--tries=num

Send a total of num probe packets per hop, defaulting to 3.

--resolve-hostnames ¶

Attempt to resolve all addresses as hostnames.

-t num ¶

--tos=num

Set type-of-service, TOS field, to num on transmitted packets.

-w num ¶

--wait=num

Set timeout in seconds, within which a returning response packet is accepted as such. Default waiting time is three seconds.

9.2 Diagnostic tokens ¶

During execution, traceroute sends three datagrams for each value for the TTL field, printing a diagnostic line of output for these. The TTL field is then steadily increased until the intended host responds, or some intermediary gateway returns a datagram to the effect that the target cannot be reached due to one reason or another.

Each line of output displays a sequence number, followed by diagnostic annotation. Any responding host has its address printed without repetition, together with a measured timing. In case there is no response within a time period of three seconds, an asterisque ‘*’ is printed.

When an intermediate router responds with an exceptional state, the time elapsed since emitting the original datagram is printed, followed by an additional short hand hint of the reason:

‘!F’

Fragmentation needed by gateway.

‘!H’

Host not reachable from gateway.

‘!N’

Network not reachable from gateway.

‘!P’

Protocol not usable at host, or within network.

‘!S’

Source routing failed at gateway.

‘!T’

Host or network not reachable for stated type of service, TOS.

‘!U’

Isolated host, not reachable.

‘!X’

Forbidden by remote administration.

whois [OPTION...] OBJECT...

10.1 Command line options ¶

-a ¶

Search all data bases.

-F ¶

Fast and raw output. Implies -r.

-g source:first-last ¶

Find updates for an object from provider source, starting from the version with serial key first, and ending with serial key last.

-h host ¶

--server=host

Connect to server host.

-H ¶

Hide legal disclaimers.

-i attr[,attr2…] ¶

Do an inverse lookup for specified attributes. Use a comma separated list for multiple attributes.

-l ¶

One level less specific lookup. Applies to RPSL only.

-L ¶

Find all less specific matches.

-m ¶

Find more specific matches, one level deeper.

-M ¶

Find all more specific matches.

-p port ¶

Connect to server port port.

-q {version|sources} ¶

Query specified server info. Applies to RPSL only.

-r ¶

Turn off recursive lookups.

-R ¶

Force output to show local copy of the domain object, even if it contains a referral.

-s source[,source2…] ¶

Search the data base at source. A comma separated list queries multiple providers.

-S ¶

Tell server to refrain from syntactic sugar.

-t type ¶

Request a template for objects of type type. Use the value ‘all’ for a list of possible types.

-T type[,type2…] ¶

Search only for objects of type type. A comma separated list allows for multiple types.

-V ¶

--verbose

Verbosely explain all actions taken.

-x ¶

Search only for exact matches. Applicable only to RPSL.

10.2 Environment variables ¶

whois holds an internal list of information servers and their assigned data bases. Queries are examined against this list to select the most plausible server, but the hint can always be overruled on the command line by use of the option -h. If neither of these have a say, then the default server to ask is ‘whois.internic.net’, but this name is in turn overruled by a server name in the environment variable WHOIS_SERVER.

LANG

When the server ‘whois.nic.ad.jp’ is queried, and only then, any non-Japanese locale in LANG will ask the server to reply with English text, not Japanese.

WHOIS_HIDE

When set, the effect on whois is as if the option -H had been given.

WHOIS_SERVER

Data base server to query when internal hinting is inconclusive. When unset, ‘whois.internic.net’ is used as default server.

ftp [option...] [host [port]]
pftp [option...] [host [port]]

ftp [option...] user@host [port]
pftp [option...] user@host [port]

Next: Commands interpreted by ftp, Up: ftp: FTP client   [Contents][Index]

11.1 Command line options ¶

Many command line options have counterparts among the commands handled by the internal interpreter.

-4 ¶

--ipv4

Initially set addressing to IPv4 only.

-6 ¶

--ipv6

Initially set addressing to IPv6 only.

-A ¶

--active

Enable active mode transfer. Default mode for ftp.

-d ¶

--debug

Enable debugging output and possibly also socket debugging.

-e ¶

--no-edit

Disables the editing of commands. This is default setting for batch mode, without a TTY, or when the environment variable TERM is not set or its value is ‘dumb’.

-g ¶

--no-glob

Disables file name globbing.

-i ¶

--no-prompt

Turns off interactive prompting during multiple file transfers.

-N netrc ¶

--netrc=netrc

Set a preferred location of the .netrc file, thus overriding any environment setting in NETRC, as well as the default location $HOME/.netrc, see The .netrc file.

-n ¶

--no-login

Restrains ftp from attempting auto-login upon initial connection. If auto-login is enabled, ftp will check the .netrc (see The .netrc file) file in the user’s home directory for an entry describing an account on the remote machine. If no entry exists, ftp will prompt for the remote machine login name (default is the user identity on the local machine), and, if necessary, prompt for a password and an account with which to login.

-p ¶

--passive

Enable passive mode transfer. Default mode when invoked as pftp.

--prompt[=prompt] ¶

Print a command-line prompt, even if not on a tty. If prompt is supplied, its value is used instead of the default ‘ftp> ’. Notice, that the argument is optional.

-t ¶

--trace

Enable packet tracing (not implemented).

-v ¶

--verbose

Start in verbose mode, printing informational messages. This is default for interactive mode.

Next: Environment variables in use, Previous: Command line options, Up: ftp: FTP client   [Contents][Index]

11.2 Commands interpreted by ftp ¶

When ftp is awaiting commands from the user, a prompt is displayed. The default string is ‘ftp>’, but it can been changed with a command line option, perhaps to enhance uniqueness while recording a session.

Be aware that correct execution of many commands depends upon a proper behavior of the remote server. The following commands are recognized by ftp itself. Command names can be abbreviated to the shortest unique string with identical beginning.

! [command [args]]

Invoke an interactive shell on the local machine. If there are arguments, the first is taken to be a command to execute directly, with the rest of the arguments as its arguments.

$ macro-name [args]

Execute the macro macro-name that was defined with the macdef command. Arguments are passed to the macro unglobbed.

account [passwd]

Supply a supplemental password required by a remote system for access to resources, once a login has been successfully completed. If no argument is included, the user will be prompted for an account password in non-echoing input mode.

append local-file [remote-file]

Append a local file to a file on the remote machine. If remote-file is left unspecified, the local file name is used in naming the remote file after being altered by any ntrans or nmap setting. File transfer uses the current settings for type, format, mode, and structure.

ascii

Set the file transfer type to network ASCII. This is the default type, except when two unices are communicating.

bell

Arrange that a bell be sounded after each file transfer command is completed.

binary

Set the file transfer type to support binary image transfer. This transfer type is selected during initial handshake, should the client on a Unix system recognize that the server is also running on a Unix system.

bye

quit

Terminate the FTP session with the remote server and exit ftp. An end of file will also terminate the session and exit.

case

Toggle the remote computer’s use of letter case mapping during mget commands. When case is ‘on’, a file name at the remote site whose every letter appear in upper case, will be renamed in such a way that all letters are changed to lower case for a local copy of the same file. The default setting is ‘off’,

cd remote-directory

Change the working directory on the remote machine to remote-directory.

cdup

Change the remote machine’s working directory to the parent of the current working directory.

chmod mode file-name

Change the access permission of the file file-name on the remote system to mode.

close

disconnect

Terminate the FTP session with the present remote server, and return to the command interpreter. Any defined macros are erased.

cr

Toggle carriage return stripping during ASCII type file retrieval. Records are denoted by a carriage return/linefeed sequence during ASCII type file transfer. When cr is ‘on’ (the default), carriage returns are stripped from this sequence to conform with the UNIX single linefeed record delimiter. Records on non-UNIX remote systems may contain single linefeeds; when an ASCII type transfer is made, these linefeeds may be distinguished from a record delimiter only when cr is ‘off’.

delete remote-file

Delete the file remote-file on the remote machine.

debug [debug-value]

Toggle debugging mode. If an optional debug-value is specified it is used to set the debugging level. When debugging is on, ftp prints each command sent to the remote machine, preceded by the string ‘-->’.

dir [remote-directory] [local-file]

Print a listing of the contents in the directory remote-directory, and, optionally, place the output in local-file. If interactive prompting is set, ftp will prompt the user to verify that the last argument is the intended local file to receive output. If no directory is specified, the current working directory on the remote machine is used. If no local file is specified, or if local-file is a dash ‘-’, then output is displayed on the terminal.

epsv4

Toggle the use of EPSV/EPRT for IPv4 addressing. Default is off.

form format

Set the file transfer form to format. The only supported format is ‘non-print’.

get remote-file [local-file]

recv remote-file [local-file]

Retrieve the remote-file and store it on the local machine. If a local file name is not specified, the local copy is given the same name as is stated for the remote original, subject to alteration by the current case, ntrans, and nmap settings. The current settings for type, form, mode, and structure are effective during file transfer.

glob

Toggle file name expansion for mdelete, mget, and mput. If globbing is turned off with glob, the file name arguments are taken literally and are not expanded. Globbing for mput is done as in csh syntax. For mdelete and mget, each remote file name is expanded separately on the remote machine and the lists are not merged. Expansion of a directory name is likely to be different from expansion of the name of an ordinary file: the exact result depends on the remote operating system and on the FTP server, and can be previewed by issuing ‘mls remote-files -’.

Note: mget and mput are not meant to transfer entire directory subtrees of files. That can be achieved by transferring an already created tar or cpio archive of the subtree, then making certain that ftp uses binary mode.

hash [size]

In the absence of an argument, toggle the state of hash-sign (‘#’) printing after each transferred data block. The optional argument selects the size of data blocks, and unconditionally activates printing. The default size is 1024 bytes. For convenience, the size can be written with postfix multipliers ’k’, ’K’, ’m’, ’M’, and ’g’, ’G’, to specify kilobytes, Megabytes, and Gigabytes, respectively.

help [command]

? [command]

Print an informative message about the meaning of command. If no argument is given, ftp prints a list of the known commands.

idle [seconds]

Set the inactivity timer on the remote server to seconds seconds. If seconds is omitted, the current inactivity timer is printed.

ipv4

Select IPv4 as the only addressing scheme.

ipv6

Select IPv6 as the only addressing scheme.

ipany

Allow IPv4 as well as IPv6 addressing.

lcd [directory]

Change the working directory on the local machine. If no directory is specified, the user’s home directory is used.

lpwd

Print the name of the current working directory on the local machine.

ls [remote-directory] [local-file]

Print a listing of the contents of a directory on the remote machine. The listing includes any system-dependent information that the server chooses to include; for example, most UNIX systems will produce output like the command ls -l does. Use nlist for a simple file listing.

If remote-directory is left unspecified, the current working directory is used. With interactive prompting set, ftp will prompt the user to verify that the last argument is indeed the intended local file for storing output. Should no local file be specified, or if local-file is a dash ‘-’, then output is sent to the terminal.

macdef macro-name

Define a macro called macro-name, with subsequent lines as the macro definition. A null line (consecutive newline characters in a file, or carriage returns at a terminal) terminates macro input mode. There is a limit of 16 macros and a total of 4096 characters shared by all defined macros. Only the first eight characters in macro-name are significant when determining which macro to execute. Macros remain defined until a close command is executed.

The macro processor interprets ‘$’ and ‘\’ as special characters. A ‘$’ followed by a number (one or more digits) is replaced by the corresponding argument on the macro’s invocation command line. A ‘$’ followed by the letter ‘i’ tells the macro processor that the macro is to perform a loop. On the first pass, ‘$i’ is replaced by the first argument on the macro’s invocation command line, while on the second pass it is replaced by the second argument, and so forth. Iteration proceeds until all arguments have been consumed.

A backslash ‘\’ followed by any character is replaced by that character. Use the backslash ‘\’ to prevent special treatment of the dollar sign ‘$’, as was just explained.

A macro can execute a macro, allowing recursion. In order to avoid exhausting the stack and thus crashing ftp, the nesting depth of macro execution is limited to a compile time constant.

mdelete [remote-files]

Delete all remote-files on the remote machine.

mdir remote-files local-file

Like dir, except multiple remote files may be specified. If interactive prompting is on, ftp will prompt the user to verify that the last argument is indeed the intended local file for storing any output from mdir.

mget remote-files

Expand the remote-files on the remote machine and execute a get for each file name thus produced. Resulting file names will then be processed according to case, ntrans, and nmap settings. Files are transferred to the local working directory, which can be changed with lcd directory; new local directories can be created with ! mkdir directory.

mkdir directory-name

Make a directory on the remote machine.

mls remote-files local-file

Like nlist, except multiple remote files may be specified, and the local-file must be specified. If interactive prompting is on, ftp will prompt the user to verify that the last argument is the intended local file for storing output. A dash ‘-’ is accepted as last argument without check!

mode [mode-name]

Set the file transfer mode to mode-name. The default mode is ‘stream’, and it is also the only implemented mode.

modtime file-name

Show the last modification time of the file on the remote machine.

mput local-files

Consider the arguments to be local names and expand any wild card. Execute a put for each file in the resulting list. The remote file names are then computed by use of ntrans and nmap settings.

newer file-name

Get the file only if the modification time of the remote file is more recent than the file on the current system. If the file does not exist on the current system, the remote file is considered newer. In other respects, this command is identical to get.

nlist [remote-directory] [local-file]

Print a list of the files in a directory on the remote machine. If remote-directory is left unspecified, the current working directory is used. If interactive prompting is on, ftp will prompt the user to verify that the last argument is the intended local file for storing output. If no local file is specified, or if local-file is ‘-’, the output is sent to the terminal.

nmap [inpattern outpattern]

Set or unset the file name mapping mechanism. If no arguments are specified, the file name mapping mechanism is unset. Name mapping is applied during mput and put commands issued without a specified remote target filename. It as also applied to local file names during mget and get commands issued without local target file name. This command is useful when connecting to a non-UNIX remote computer with different file naming conventions or practices.

The mapping follows the pattern set by inpattern and outpattern. The template inpattern is used on incoming filenames (which may have already been processed according to the ntrans and case settings). Variable templating is accomplished by including the sequences ‘$1’, ‘$2’, …, ‘$9’ in inpattern. Use ‘\’ to prevent this special treatment of the character ‘$’. All other characters are treated literally, and must be matched in a file name for inpattern to bind substrings to variables.

For example, take a pattern ‘$1.$2’ and a file name mydata.data. Then ‘$1’ would have the value ‘mydata’, and ‘$2’ would be ‘data’.

outpattern determines the final file name. The sequences ‘$1’ to ‘$9’ are replaced by any values bound to them by inpattern. A special sequence ‘$0’ always contains the original filename. In addition, a bracketted sequence ‘[seq1,seq2]’ expands to seq1 if seq1 contains a non-empty string, and expands to seq2 otherwise. For example, the command

nmap $1.$2.$3 [$1,$2].[$2,file]

would yield the output file name myfile.data for input names myfile.data and myfile.data.old, but produces myfile.file from the input myfile, and myfile.myfile from .myfile.

Spaces may be included in outpattern, but are easily removed:

nmap $1 |sed "s/ *$//" > $1

Use a backslash ‘\’ to escape the characters ‘$’, ‘[’, ‘]’, and ‘,’.

ntrans [inchars [outchars]]

Set or unset the filename character translation mechanism. If no arguments are specified, the filename character translation mechanism is unset. If arguments are specified, characters in remote filenames are translated during mput commands and put commands issued without a specified remote target filename. If arguments are specified, characters in local filenames are translated during mget commands and get commands issued without a specified local target filename. This command is useful when connecting to a non-UNIX remote computer with different file naming conventions or practices.

Characters in a filename matching a character in inchars are replaced with the corresponding character in outchars. If the character’s position in inchars is longer than the length of outchars, the character is deleted from the file name.

open host [port]

open user@host [port]

Establish a connection to the specified FTP server at host. An optional port number may be supplied, in which case, ftp will attempt to contact the server at that specific TCP port. If the autologin option is on (is so by default), ftp will also attempt to automatically log the user in to the FTP server.

The second form of invocation sets the remote user name to user, which otherwise is taken as identical to the user identity owning the local session.

passive

Toggle passive mode. If passive mode is turned on (default is off), the ftp client will send a PASV command for all data connections instead of the usual PORT command. The PASV command requests that the remote server open a port for the data connection and return the address of that port. The remote server listens on that port and the client connects to it. When using the more traditional PORT command, the client listens on a port and sends that address to the remote server, who connects back to it. Passive mode is useful when using ftp through a gateway router or host that controls the directionality of traffic. (Note that though ftp servers are required to support the PASV command by RFC 1123, some do not.) If epsv4 has been set to on, the client will attempt EPSV before PASV for IPv4. As a last resort LPSV is attempted. With IPv6 only EPSV and LPSV are possible.

prompt

Toggle interactive prompting. Interactive prompting occurs during multiple file transfers to allow the user to selectively retrieve or store files. If prompting is turned off (default is on), any mget or mput will transfer all files, and any mdelete will delete all files.

proxy ftp-command

Execute an ftp command on a secondary control connection. This command allows simultaneous connection to two remote FTP servers for transferring files between the two servers. The first proxy command should be open, to establish the secondary control connection. Enter the command proxy ? to see other commands usable for the secondary connection. The following commands behave differently when prefaced by proxy: open will not define new macros during the auto-login process, close will not erase existing macro definitions, get and mget transfer files from the host on the primary control connection to the host on the secondary control connection, and put, mput, and append transfer files from the host on the secondary control connection to the host on the primary control connection.

Note that the protocol command PASV must be understood by the server on the secondary control connection for this kind of file transfer to succeed.

put local-file [remote-file]

send local-file [remote-file]

Store a local file on the remote machine. If remote-file is left unspecified, the local file name is used after processing according to any ntrans or nmap settings in naming the remote file. File transfer uses the current settings for type, format, mode, and structure.

pwd

Print the name of the current working directory on the remote machine.

quote arg…

The arguments specified are sent, verbatim, to the remote FTP server.

reget remote-file [local-file]

reget acts like get, except that if local-file exists and is smaller than remote-file, then local-file is presumed to be a partially transferred copy of remote-file and the transfer is continued from the apparent point of failure. This command is useful when transferring very large files over networks that are prone to dropping connections.

rhelp [command-name]

Request help from the remote FTP server. If command-name is specified it is passed to the server as well.

rstatus [file-name]

With no arguments, show status of remote machine. If filename is specified, show status of file-name on remote machine.

rename [from] [to]

Rename the file from on the remote machine as to. Name mapping takes effect without to.

reset

Clear reply queue. This command re-synchronizes command/reply sequencing with the remote FTP server. Resynchronization may be necessary following a violation of the FTP protocol by the remote server.

restart marker

Restart the immediately following get or put at the indicated marker. On UNIX systems, marker is usually a byte offset into the file.

rmdir directory-name

Delete a directory on the remote machine.

runique

Toggle the storing of files on the local system with unique filenames. If a file already exists with a name equal to the intended local file name for a get or mget command, then a string ‘.1’ is appended to the name. If the resulting name matches another existing file, ‘.2’ is appended to the original name. If this process continues up to ‘.99’, an error message is printed, and the transfer does not take place. The generated unique filename will be reported. Note that runique will not affect local files generated from a shell command. The default value is off.

sendport

Toggle the use of PORT commands. By default, ftp will attempt to use a PORT command when establishing a connection for each data transfer. The use of PORT commands can prevent delays when performing multiple file transfers. If the PORT command fails, ftp will use the default data port. When the use of PORT commands is disabled, no attempt will be made to use PORT commands for each data transfer. This is useful for certain FTP implementations which do ignore PORT commands but, incorrectly, indicate they’ve been accepted.

site arg…

The arguments specified are sent, verbatim, to the remote FTP server as a SITE command.

size file-name

Return size of file-name on remote machine.

status

Show the current status of ftp.

struct [struct-name]

Set the file transfer structure to struct-name. By default ‘file’ structure is used, which also is the only supported value.

sunique

Toggle storing of files on remote machine under unique file names. Remote FTP server must support FTP protocol STOU command for successful completion. The remote server will report unique name. Default value is off.

system

Show the type of operating system running on the remote machine.

tenex

Set the file transfer type to that needed to talk to TENEX machines.

trace

Toggle packet tracing (feature is not implemented).

type [type-name]

Set the file transfer type to type-name. If no type is specified, the current type is printed. The recognized type names are ‘ascii’, ‘binary’, ‘ebcdic’, ‘image’, and ‘tenex’. The default type is network ASCII.

umask [newmask]

Set the default umask on the remote server to newmask. If newmask is omitted, the current umask is printed.

user user-name [password] [account]

Identify yourself to the remote FTP server. If the password is not specified and the server requires it, ftp will prompt the user for it (after disabling local echo). If an account field is not specified, and the FTP server requires it, the user will be prompted for it. If an account field is specified, an account command will be relayed to the remote server after the login sequence is completed if the remote server did not require it for logging in. Unless ftp is invoked with auto-login disabled, this process is done automatically on initial connection to the FTP server.

verbose

Toggle verbose mode. In verbose mode, all responses from the FTP server are displayed to the user. In addition, if verbose is on, when a file transfer completes, statistics regarding the efficiency of the transfer are reported. By default, verbose is on.

Command arguments which have embedded spaces may be inclosed within citation characters ‘"’.

Next: Aborting a file transfer, Previous: Commands interpreted by ftp, Up: ftp: FTP client   [Contents][Index]

11.3 Environment variables in use ¶

ftp accesses the following environment variables.

HOME

Used for locating a .netrc file, if one exists.

NETRC

Alternate location of the .netrc file, taking precedence over the standard location.

SHELL

For determining the default shell interpreter.

Next: File naming conventions, Previous: Environment variables in use, Up: ftp: FTP client   [Contents][Index]

11.4 Aborting a file transfer ¶

To abort a file transfer, use the terminal interrupt key (usually C-c). Sending transfers will be immediately halted. Receiving transfers will be halted by sending a FTP protocol command ABOR to the remote server, discarding any further data received. The speed at which this is accomplished depends upon the remote server’s support for ABOR processing. If the remote server does not support the ABOR command, an ‘ftp>’ prompt will not appear until the remote server has completed sending the requested file.

The terminal interrupt key sequence will be ignored when ftp has completed any local processing and is awaiting a reply from the remote server. A long delay in this mode may result from the ABOR processing described above, or from unexpected behavior by the remote server, including violations of the FTP protocol. If the delay results from unexpected remote server behavior, the local ftp program must be killed by hand.

Next: File transfer parameters, Previous: Aborting a file transfer, Up: ftp: FTP client   [Contents][Index]

11.5 File naming conventions ¶

Files specified as arguments to ftp commands are processed according to the following rules.

1. If the file name ‘-’ is specified, standard input (for reading) or standard output (for writing) is used.
2. If the first character of the file name is ‘|’, the remainder of the argument is interpreted as a shell command. ftp then forks a shell, using popen with the argument supplied, and reads/writes from standard input/output. If the shell command includes spaces, the argument must be quoted; e.g. ‘"ls -lt"’.

   A particularly useful example of this mechanism in action, is

   which allows the user to scroll through a long directory listing.

3. Failing the above checks, if globbing is enabled, local file names are expanded according to the rules used by csh; c.f. the glob command. If the ftp command expects a single local file (e.g. put), only the first filename generated by the globbing operation is used.
4. For the commands mget and get with unspecified local file name, the local file name is set to the remote file name, which may be altered by a case, ntrans, or nmap settings. The resulting file name may then be modified if runique is set.
5. For the commands mput and put with unspecified remote file name, the remote file name is copied from the local file name, which may be altered by a ntrans or nmap settings. The resulting file name may also be modified by the remote server if sunique is set.

Next: The .netrc file, Previous: File naming conventions, Up: ftp: FTP client   [Contents][Index]

11.6 File transfer parameters ¶

The FTP specification includes many parameters which may affect a file transfer. The type may be one of ‘ascii’, ‘image’ (binary), ‘ebcdic’, and ‘local’ byte size (for PDP-10’s and PDP-20’s mostly). ftp supports the ‘ascii’ and ‘image’ types of file transfer, plus local byte size 8 for tenex mode transfers.

ftp supports only the default values for the remaining file transfer parameters: mode, form, and struct.

An error in the treatment of carriage returns in the 4.2BSD ascii-mode transfer code has been corrected by the present implementation. This correction may result in corrupt transfers of binary files to and from 4.2BSD servers, when done using the ascii type. Avoid this problem by using the binary image type.

Previous: File transfer parameters, Up: ftp: FTP client   [Contents][Index]

11.7 The .netrc file ¶

The .netrc file contains login and initialization information used by the auto-login process. It generally resides in the user’s home directory, but a location outside of the home directory can be set using the environment variable NETRC. Both locations are overridden by the command line option -N. The selected file must be a regular file, or access will be denied.

The following tokens are recognized; they may be separated by spaces, tabs, or new-lines:

‘machine name’

Identify a remote machine name. The auto-login process searches the .netrc file for a machine token that matches the remote machine specified on the ftp command line or as an open command argument. Once a match is made, the subsequent .netrc tokens are processed, stopping when the end of file is reached or another machine or a default token is encountered.

‘default’

This is the same as machine name except that default matches any name. There can be only one default token, and it must be after all machine tokens. This is normally used as:

default login anonymous password user@site

thereby giving the user automatic anonymous ftp login to machines not specified in .netrc. This can be overridden by using the -n flag to disable auto-login.

‘login name’

Identify a user on the remote machine. If this token is present, the auto-login process will initiate a login using the specified name.

‘password string’

Supply a password. If this token is present, the auto-login process will supply the specified string if the remote server requires a password as part of the login process. Note that if this token is present in the .netrc file for any user other than anonymous, ftp will abort the auto-login process if the .netrc is readable by anyone besides the user.

‘account string’

Supply an additional account password. If this token is present, the auto-login process will supply the specified string if the remote server requires an additional account password, or the auto-login process will initiate an ACCT command if it does not.

‘macdef name’

Define a macro. This token functions like the ftp macdef command functions. A macro is defined with the specified name; its contents begin with the next .netrc line and continue until a null line (consecutive new-line characters) is encountered. If a macro named init is defined, it is automatically executed as the last step in the auto-login process.

rcp [option]... old-file new-file
rcp [option]... files... directory

12.1 Command line options ¶

-4 ¶

--ipv4

Use only IPv4.

-6 ¶

--ipv6

Use only IPv6.

-d directory ¶

--target-directory=directory

Copy all source arguments into directory.

-f ¶

--from

(Server mode only.) Copying from remote host.

-k realm ¶

--realm=realm

The option requests rcp to obtain tickets for the remote host in realm realm instead of the remote host’s realm.

-K ¶

--kerberos

Turns off all Kerberos authentication.

-p ¶

--preserve

Causes rcp to attempt to preserve (duplicate) in its copies the modification times and modes of the source files, ignoring the umask. By default, the mode and owner of the target file are preserved if the target itself already exists; otherwise the mode of the source file is modified by the umask setting on the destination host.

-r ¶

--recursive

If any of the source files are directories, rcp copies each subtree rooted at that name; in this case the destination must be a directory.

-t ¶

--to

(Server mode only.) Copying to remote host.

-x ¶

--encrypt

Turns on encryption for all data passed via the rcp session. This may impact response time and CPU utilization, but provides increased security.

rcp doesn’t detect all cases where the target of a copy might be a file in cases where only a directory should be legal.

rcp can be confused by any output generated by commands in a .login, .profile, or .cshrc file on the remote host.

The destination user and hostname may have to be specified as ‘rhost.rname’ when the destination machine is running the 4.2BSD version of rcp.

rexec --user=login --password=pass --host=host  \
      [OPTION] command

13.1 Command line options ¶

-4 ¶

--ipv4

Use only IPv4 connections as all times.

-6 ¶

--ipv6

Use only IPv6 connections.

-a ¶

--ipany

Allow any address family for connections. This is the default.

-e ¶

--error=port

Specify the TCP port to use for stderr redirection, in case it is not specified a random port will be used.

-h

--host=name ¶

Specify the host with whom to connect: symbolic name or address.

-n ¶

--noerr

If specified, an error stream will not be created.

-p ¶

--password=passwd

Specify the password for logging-in. The special value consisting of a single dash ‘-’ will make rexec read a single line from stdin. This input is then used as password and is passed as such to the remote server. Thus it is possible to hide vital access information slightly better than the full disclosure implicit in the text of a command line option.

-P ¶

--port=num

Specify to which numerical port a connection shall be sought. If it is not specified, then use port 512/tcp by default.

-u ¶

--user=name

Specify the user with whom to log into the server.

# ln -s /usr/bin/rlogin hostname
# hostname -8

14.1 Command line options ¶

The options are as follows :

-4 ¶

--ipv4

Use only IPv4.

-6 ¶

--ipv6

Use only IPv6.

-8 ¶

--8-bit

Allows an eight-bit input data path at all times; otherwise parity bits are stripped except when the remote side’s stop and start characters are other than C-S/C-Q.

-d ¶

--debug

Turns on socket debugging on the TCP sockets used for communication with the remote host.

-e char ¶

--escape=char

Allows user specification of the escape character, which is ‘~’ by default. This specification may be as a literal character, or as an octal value in the form ‘\nnn’.

-E ¶

--no-escape

Stops any character from being recognized as an escape character. When used with the -8 option, this provides a completely transparent connection.

-l user ¶

--user=user

By default, the remote username is the same as the local username. This option, and the ‘user@host’ format, allow the remote user name to be made explicit, or changed.

The next three options are available only if the program has been compiled with support for Kerberos authentication.

-k realm ¶

--realm=realm

The option requests rlogin to obtain tickets for the remote host in realm realm instead of the remote host’s realm.

-K ¶

--kerberos

Turns off all Kerberos authentication.

-x ¶

--encrypt

Turns on encryption for all data passed via the rlogin session. This may impact response time and CPU utilization, but provides increased security.

14.2 Escape characters and flow control ¶

As long as the connection stands, the client program rsh is observing the input stream in order to detect so called escape sequences, allowing the user to execute some local actions without having to tear down the remote connection.

The sequences consist of two characters, the first of which always is the distinguished character escape-char. The following sequences are supported:

• escape-char . disconnects from the remote host.
• escape-char C-d does the same. (Termios character ‘VEOF’.)
• escape-char C-z suspends the session, halting the remote process, but keeping it ready for resumed processing. The user is given access to a local shell. (Termios character ‘VSUSP’.)
• escape-char delayed-suspend-char implements a half-way suspend, in the sense of stopping local input from reaching the remote side, but still displaying all output from the remote host on the local terminal. The remote process is still running, but the local user is given a local shell until the resuming the original command. Normally, only BSD systems offer this mode. (Termios character ‘VDSUSP’.)

By default, the character tilde ‘~’ is assigned to escape-char, but it can be changed using the option --escape. The processing of escape sequences can even be disable using the option --no-escape. On BSD systems, delayed-suspend-char is usually set to C-Y. It displays as ‘dsusp’ using stty.

All echoing takes place at the remote site, so that the rlogin is transparent except possibly for transmission delays. Flow control via C-S and C-Q, if at all supported, will stop and start the flow of data on the local terminal. Flushing of input and output on interrupts is also handled properly.

On the server side the iruserok and ruserok functions are used to authenticate the connection request, unless Kerberised mode is in effect. See the appropriate man pages for more information.

14.3 Kerberos Authentication ¶

If rlogin was compiled with kerberos support, options -x, -k, -K are available. Each user may have a private authorization list in the file .k5login in their home directory. Each line in this file should contain a Kerberos principal name of the form ‘principal/instance@realm’. If the originating user is authenticated to one of the principals named in .k5login, access is granted to the account. The principal ‘accountname@localrealm’ is granted access if there is no .k5login file. Otherwise a login and password will be prompted for on the remote machine as in login. To avoid certain security problems, the .k5login file must be owned by the remote user. If Kerberos authentication fails, a warning message is printed and the standard Berkeley rlogin is used instead.

# ln -s /usr/bin/rsh hostname
# hostname ls

15.1 Command line options ¶

The options are as follows :

-4 ¶

--ipv4

Use only IPv4.

-6 ¶

--ipv6

Use only IPv6.

-d ¶

--debug

Turns on socket debugging used for communication with the remote host.

-l user ¶

--user=user

By default, the remote username is the same as the local username. The -l option and the ‘username@host’ format allow the remote user name to be specified. Kerberos authentication is used, whenever available, and authorization is determined as in rlogin (see rlogin: Remote login).

-n ¶

--no-input

Use /dev/null for all input, telling the server side that we send no material. This can prevent the remote process from blocking, should it optionally accept more input. The option is void together with encryption.

The next three options are available only if the program has been compiled with support for Kerberos authentication.

-k realm ¶

--realm=realm

The option requests rsh to obtain tickets for the remote host in realm realm instead of the remote host’s realm.

-K ¶

--kerberos

Turns off all Kerberos authentication.

-x ¶

--encrypt

Turns on encryption for all data passed via the rsh session. This may impact response time and CPU utilization, but provides increased security.

Finally, some compatibility options are present:

-8

--8-bit

-e char

--escape=char

-E

--no-escape

Ignored during normal operation, but passed on to rlogin when rsh is invoked without a command argument.

15.2 Note on stream redirections ¶

Beware that non-quoted shell metacharacters are interpreted on the local machine, while quoted metacharacters are interpreted on the remote machine. For example:

rsh otherhost  cat remotefile >> localfile
rsh otherhost  cat remotefile ">>" otherfile

The first command appends the contents of remotefile, as found on the remote host, to the file localfile on the local host, since the local shell will intercept the redirection and will thus receive whatever the remote process directs to stdout.

In contrast, the second command will append the contents of the same file remotefile to a file named otherfile again, but this time the file is located on the remote host. The effect of quoting the redirection operator is to execute the command

cat remotefile >> localfile

entirely on the remote most, whence stdout at the remote host will have nothing to transmit to the listening local host!.

16.1 Invoking ¶

The command line arguments are as follows:

person

If you wish to talk to someone on your own machine, then person is just the other person’s login name. If you wish to talk to a user on another host, then person is of the form ‘user@host’.

ttyname

If you wish to talk to a local user who is logged in more than once, the argument ttyname may be used to indicate the appropriate terminal name, where ttyname typically is of the form ‘ttyXX’, or ‘pts/X’.

When first called, talk sends a message to the addressed user:

Message from TalkDaemon@his_machine...
talk: connection requested by your_name@your_machine.
talk: respond with: talk your_name@your_machine

At this point, the recipient of the message could elect to accept the call and to establish a connection by typing:

talk your_name@your_machine

It doesn’t matter from which machine the recipient replies, as long as his login-name is the same. Once communication is established, the two parties may type text simultaneously, with their output appearing in separate windows. Typing C-L will cause the screen to be reprinted, while erase, kill, and word kill characters will behave normally. In addition, C-D will cause both windows to be locally cleared of all text. This keystroke will appear as a simple ‘^D’ on the remote terminal, though. It signals to the other party that you yourself have just cleared your terminal of all text.

To exit, just type an interrupt character C-C; talk then moves the cursor to the bottom of the screen and restores the terminal to its previous state.

The ability to talk may be enabled or disabled by use of the mesg command. It is system dependent whether this message passing is enabled at the outset of a terminal session. Certain commands, in particular nroff and pr, disable messages in order to prevent messy output.

telnet [OPTION...] [HOST [PORT]]

17.1 Command line options ¶

-4 ¶

--ipv4

Use only IPv4.

-6 ¶

--ipv6

Use only IPv6.

-8 ¶

--binary

Use an 8-bit data path.

-a ¶

--login

Attempt automatic login.

-b address ¶

--bind=address

Bind to specific local address.

-c ¶

--no-rc

Do not read the user’s file $HOME/.telnetrc.

-d ¶

--debug

Turn on socket level debugging.

-e char ¶

--escape=char

Use char as escape character.

-E ¶

--no-escape

Do not use an escape character.

-k realm ¶

--realm=realm

Request Kerberos realm realm instead of whatever is declared as default realm in the system’s or user’s settings.

-K ¶

--no-login

Do not automatically login to the remote system.

-l user ¶

--user=user

Attempt automatic login as user.

-L ¶

--binary-output

Use an 8-bit data path for output only.

-n file ¶

--trace=file

Record trace information into file.

-r ¶

--rlogin

Display a user-interface similar to that of rlogin.

-x ¶

--encrypt

If possible, encrypt the data stream.

-X atype ¶

--disable-auth=atype

Disable authentication of type atype. Use this option multiple times if more than one type is to be disabled. Standard choices are ‘null’, ‘kerberos_v4’, and ‘kerberos_v5’.

18.1 Commands ¶

Once tftp is running, it issues the prompt and recognizes the following commands:

? command-name

Print help information.

ascii

Shorthand for mode ascii

binary

Shorthand for mode binary

connect host-name [port]

Set the host (and optionally port) for transfers. Note that the TFTP protocol, unlike the FTP protocol, does not maintain connections between transfers; thus, the connect command does not actually create a connection, but merely remembers what host is to be used for transfers. You do not have to use the connect command; the remote host can be specified as part of the get or put commands.

get file-name

get remotename localname

get file…

Get a file, or a set of files, from the specified sources. The source can be in one of two forms: a file name on the remote host, if the host has already been specified, or a string of the form ‘host:filename’ to specify both a host and file name at the same time. If the latter form is used, the last hostname specified becomes the default for future transfers. When specifying a numeric IPv6 address as host part, then this address must be enclosed between square brackets, since it contains colons and would interfere with the delimiter before the file name. Brackets are optional for IPv4 addresses.

tftp> get [2001:1234::12]:issue

mode transfer-mode

Set the mode for transfers; transfer-mode may be one of ‘ascii’ or ‘binary’. The default is ‘ascii’.

put file

put localfile remotefile

put file… remote-directory

Put a file or set of files to the specified remote file or directory. The destination can be in one of two forms: a filename on the remote host, if the host has already been specified, or a string of the form ‘host:filename’ to specify both a host and filename at the same time. If the latter form is used, the hostname specified becomes the default for future transfers. If the remote-directory form is used, the remote host is assumed to be a UNIX machine. The same use of square brackets for enclosing numeric IPv6 addresses applies here, as was mentioned for the command get.

quit

Exit tftp. An end of file also exits.

rexmt retransmission-timeout

Set the per-packet retransmission timeout, in seconds.

status

Show current status.

timeout total-transmission-timeout

Set the total transmission timeout, in seconds.

trace

Toggle packet tracing.

verbose

Toggle verbose mode.

Because there is no user-login or validation within the tftp protocol, the remote site will probably have some sort of file-access restrictions in place. The exact methods are specific to each site and therefore difficult to document here.

Next: Configuration file, Up: inetd: Internet super-server   [Contents][Index]

19.1 Invocation ¶

Normally, inetd is invoked without any arguments. It does, however, support several command line options. These are:

-d ¶

--debug

Turns on debugging. With this option, inetd stays in foreground and prints additional debugging information of stderr.

--environment ¶

Pass local and remote socket information in environment variables. See Inetd Environment.

-p[file] ¶

--pidfile[=file]

Use file as location to store process ID of the running server process, thus overriding the default location. Setting an empty argument will disable the use of a file for storing the process ID.

--resolve ¶

Resolve IP addresses when setting environment variables. See Inetd Environment.

-R rate ¶

--rate=rate

Specify the maximum number of times a service can be invoked in one minute; the default is 1000.

Next: Built-in services, Previous: Invocation, Up: inetd: Internet super-server   [Contents][Index]

19.2 Configuration file ¶

Upon execution, inetd reads its configuration information from configuration files and directories named on the command line. By default these are /etc/inetd.conf and /etc/inetd.d. If the configuration pathname is a directory, all files in the directory are read and interpreted like a configuration file. All of the configuration files are read and the results are merged.

There must be an entry for each field in the configuration file, with entries for each field separated by a tab or a space. Comments are denoted by a “#” at the beginning of a line. The available fields of the configuration file are summarized in the table below (optional parts are enclosed in square brackets):

[service node:]service name

The service-name entry is the name of a valid service in the file /etc/services. For “internal” services (see Built-in services), the service name must be the official name of the service (that is, the first entry in /etc/services), or a numeric representation thereof. For TCPMUX services, the value of the ‘service name’ field consists of the string ‘tcpmux’ followed by a slash and the locally-chosen service name (see TCPMUX).

An optional ‘service node’ prefix is allowed for internet services. When present, it supplies the local addresses inetd should use when listening for that service. ‘Service node’ consists of a comma-separated list of addresses. Both symbolic host names and numeric IP addresses are allowed. Symbolic hostnames are looked up in DNS service. If a hostname has multiple address mappings, inetd creates a socket to listen on each address.

To avoid repeating an address that occurs frequently, a line with a host address specifier and colon, but no further fields is allowed, e.g.:

The address specifier from such a line is remembered and used for all further lines lacking an explicit host specifier. Such a default address remains in effect until another such line or end of the configuration is encountered, whichever occurs first.

A special hostname ‘*’ stands for the wildcard address. When used in a normal configuration line, it causes the default address specifier to be ignored for that line. When used in a default address specification, e.g.:

it causes any previous default address specifier to be forgotten.

socket type

The socket type should be one of ‘stream’, ‘dgram’, ‘raw’, ‘rdm’, or ‘seqpacket’, depending on whether the socket is a stream, datagram, raw, reliably delivered message, or sequenced packet socket. TCPMUX services must use ‘stream’.

protocol

The protocol must be a valid protocol as given in /etc/protocols. Examples might be ‘tcp’ or ‘udp’. TCPMUX services must use ‘tcp’. If IPv6 support is enabled the sockets will accept both IPv4 and IPv6 connections if that is supported by the OS. If inetd should only accept IPv4 or IPv6 connections, add ‘4’ or ‘6’ to the protocol name. For example ‘tcp4’ will only accept IPv4 tcp connections and ‘udp6’ will only accept IPv6 udp connections.

wait/nowait[.max]

The ‘wait/nowait’ entry specifies whether the server that is invoked by inetd will take over the socket associated with the service access point, and thus whether inetd should wait for the server to exit before listening for new service requests. Datagram servers must use ‘wait’, as they are always invoked with the original datagram socket bound to the specified service address. These servers must read at least one datagram from the socket before exiting. If a datagram server connects to its peer, freeing the socket so inetd can received further messages on the socket, it is said to be a “multi-threaded” server; it should read one datagram from the socket and create a new socket connected to the peer. It should fork, and the parent should then exit to allow inetd to check for new service requests to spawn new servers. Datagram servers which process all incoming datagrams on a socket and eventually time out are said to be “single-threaded”. comsat and talkd are both examples of the latter type of datagram server. tftpd is an example of a multi-threaded datagram server.

Servers using stream sockets generally are multi-threaded and use the ‘nowait’ entry. Connection requests for these services are accepted by inetd, and the server is given only the newly-accepted socket connected to a client of the service. Most stream-based services and all TCPMUX services operate in this manner. For such services, the number of running instances of the server can be limited by specifying optional ‘max’ suffix (a decimal number), e.g.: ‘nowait.15’.

Stream-based servers that use ‘wait’ are started with the listening service socket, and must accept at least one connection request before exiting. Such a server would normally accept and process incoming connection requests until a timeout. Other services must use ‘nowait’.

user

The user entry should contain the user name of the user as whom the server should run. This allows for servers to be given less permission than root. An optional form includes also a group name as a suffix, separated from the user name by colon or a period, i.e., ‘user:group’ or ‘user.group’.

server program

The server-program entry should contain the pathname of the program which is to be executed by inetd when a request is found on its socket. If inetd provides this service internally, this entry should be ‘internal’.

It is common usage to specify /usr/sbin/tcpd in this field.

server program arguments

The server program arguments should be just as arguments normally are, starting with argv[0], which is the name of the program. If the service is provided internally, this entry must contain the word ‘internal’, or be empty.

Next: TCPMUX, Previous: Configuration file, Up: inetd: Internet super-server   [Contents][Index]

19.3 Built-in services ¶

The inetd program provides several “trivial” services internally by use of routines within itself. All these services can operate both in ‘stream’ and in ‘dgram’ mode. They are:

echo

Send back to the originating source any data received from it. This is a debugging and measurement tool.

discard

Silently throw away any data received.

chargen

This is a character generator service. It can be operated as both stream or dgram service. When operating in ‘stream’ mode, once a connection is established a stream of data is sent out the connection (and any data received is thrown away). This continues until the calling user terminates the connection. When operating in ‘dgram’ mode, inetd listens for UDP datagrams, and for each received datagram, answers with a datagram containing a random number (between 0 and 512) of characters. Any data in the received datagram are ignored.

daytime

Send back the current date and time in a human readable form. Any input is discarded.

time

Send back the current date and time as a 32-bit integer number, nrepresenting the number of seconds since midnight, January 1, 1900.

Next: Inetd Environment, Previous: Built-in services, Up: inetd: Internet super-server   [Contents][Index]

19.4 TCPMUX ¶

The TCPMUX protocol.

A TCP client connects to a foreign host on TCP port 1. It sends the service name followed by a carriage-return line-feed <CRLF>. The service name is never case sensitive. The server replies with a single character indicating positive (+) or negative (-) acknowledgment, immediately followed by an optional message of explanation, terminated with a <CRLF>. If the reply was positive, the selected protocol begins; otherwise the connection is closed.” The program is passed the TCP connection as file descriptors 0 and 1.

If the TCPMUX service name begins with a “+”, inetd returns the positive reply for the program. This allows you to invoke programs that use stdin/stdout without putting any special server code in them.

The special service name ‘help’ causes inetd to list TCPMUX services in inetd.conf.

To define TCPMUX services, the configuration file must contain a ‘tcpmux internal’ definition.

Here are several example service entries for the various types of services:

ftp           stream  tcp   nowait root  /usr/libexec/ftpd     ftpd -l
ntalk         dgram   udp   wait   nobody:tty /usr/libexec/talkd talkd
tcpmux        stream  tcp   nowait root  internal
tcpmux/+date  stream  tcp   nowait guest /bin/date             date
tcpmux/phonebook stream tcp nowait guest /usr/bin/phonebook    phonebook

Next: Error Messages, Previous: TCPMUX, Up: inetd: Internet super-server   [Contents][Index]

19.5 Inetd Environment ¶

If a connection is made with a streaming protocol (‘stream’) and if --environment option has been given, inetd will set the following environment variables before starting the program:

PROTO

Always ‘TCP’.

TCPLOCALIP

Local IP address of the interface which accepted the connection.

TCPLOCALPORT

Port number on which the TCP connection was established.

TCPREMOTEIP

IP address of the remote client.

TCPREMOTEPORT

Port number on the client side of the TCP connection.

Additionally, if given the --remote option, inetd sets the following environment variables:

TCPLOCALHOST

DNS name of TCPLOCALIP.

TCPREMOTEHOST

DNS name of TCPREMOTEIP.

Previous: Inetd Environment, Up: inetd: Internet super-server   [Contents][Index]

19.6 Error Messages ¶

The inetd server logs error messages using syslog. Important error messages and their explanations are:

‘service/protocol server failing (looping), service terminated.’

The number of requests for the specified service in the past minute exceeded the limit. The limit exists to prevent a broken program or a malicious user from swamping the system. This message may occur for several reasons:

1. there are lots of hosts requesting the service within a short time period,
2. a “broken” client program is requesting the service too frequently,
3. a malicious user is running a program to invoke the service in a “denial of service” attack,
4. the invoked service program has an error that causes clients to retry quickly.

Use the -R option, as described above, to change the rate limit. Once the limit is reached, the service will be re-enabled automatically in 10 minutes.

‘service/protocol: No such user 'user', service ignored’

‘service/protocol: getpwnam: user: No such user’

No entry for user exists in the passwd file. The first message occurs when inetd (re)reads the configuration file. The second message occurs when the service is invoked.

‘service/protocol: No such user 'user', service ignored’

‘service/protocol: getpwnam: user: No such user’

No entry for user exists in the passwd file. The first message occurs when inetd (re)reads the configuration file. The second message occurs when the service is invoked.

‘service: can't set uid number’

‘service: can't set gid number’

The user or group ID for the entry’s user is invalid.

20.1 Configuration file ¶

syslogd reads its configuration file when it starts up and whenever it receives a hangup signal. The syslog.conf file is the main configuration file for the syslogd program. In addition, the server looks below the directory syslog.d/ for further configuration files, making it easy to arrange a common set of logging conventions in syslog.conf, augmented by system and service specific drop-in configuration in syslog.d/.

Each configuration file consists of lines with two fields: a selector field which specifies the types of messages and priorities to which the line applies, and an action field which specifies the action to be taken if a message syslogd receives matches the selection criteria. The selector field is separated from the action field by one or more tab or space characters. A rule can be split in several lines if all lines except the last are terminated with a backslash ‘\’.

There are two exceptional forms of line content. The first is the tagged selector, and the second is a comment. The latter begins with an octothorp (‘#’), also called hash, and continues until end-of-line.

A tagged selector commences with an exclamation mark, as in ‘!name’, or with a shebang, like ‘#! name’, and continues with a program name, a tag in the sense used by logger. It has the effect of applying the following configuration rules only to messages submitted with the specified tag. This selection remains in effect until another tag is selected, or until it is reset by means of stating the program name as an asterisque ‘*’.

The selector fields are encoded as a facility, followed by a period (‘.’), and a level, with no intervening white-space. The facility as well as the level are case insensitive.

The facility describes the part of the system generating the message, and is one of the following keywords: ‘auth’, ‘authpriv’, ‘cron’, ‘daemon’, ‘ftp’, ‘kern’, ‘lpr’, ‘mail’, ‘mark’, ‘news’, ‘syslog’, ‘user’, ‘uucp’ and ‘local0’ through ‘local7’. These keywords (with the exception of ‘mark’) correspond to the similar ‘LOG_’ values specified to the ‘openlog’ and ‘syslog’ library routines. See Syslog in The GNU C Library Reference Manual, for details.

The level describes the severity of the message, and is a keyword from the following ordered list (higher to lower): ‘emerg’, ‘alert’, ‘crit’, ‘err, warning’, ‘notice’, ‘info’ and ‘debug’. These keywords correspond to the similar ‘LOG_’ values specified to the syslog library routine.

See syslog and vsyslog in The GNU C Library Reference Manual, for a further descriptions of both the facility and level keywords and their significance.

If a received message matches the specified facility and is of the specified level (or a higher level), the action specified in the action field will be taken.

Multiple selectors may be specified for a single action by separating them with semicolon (‘;’) characters. It is important to note, however, that each selector can modify the ones preceding it.

Multiple facilities may be specified for a single level by separating them with comma (‘,’) characters.

An asterisk (‘*’) can be used to specify all facilities or all levels. Two asterisks (‘**’) specifie all facilities not named previously in the configuration file.

By default, a level applies to all messages with the same or higher level. The equal (‘=’) character can be prepended to a level to restrict this line of the configuration file to messages with the very same level.

An exclamation mark (‘!’) prepended to a level or the asterisk means that this line of the configuration file does not apply to the specified level (and higher ones). In conjunction with the equal sign, you can exclude single levels as well.

The special facility ‘mark’ receives a message at priority ‘info’ every 20 minutes. This is not enabled by a facility field containing an asterisk.

The special level ‘none’ disables a particular facility.

The action field of each line specifies the action to be taken when the selector field selects a message. There are five forms:

• A pathname (beginning with a leading slash). Selected messages are appended to the file.

  You may prepend a minus (‘-’) to the path to omit syncing the file after each message log. This can cause data loss at system crashes, but increases performance for programs which use logging extensively.

• A named pipe, beginning with a vertical bar (‘|’) followed by a pathname. The pipe must be created with mkfifo before syslogd reads its configuration file. This feature is especially useful for debugging.
• A hostname (preceded by an at (‘@’) sign). Selected messages are forwarded to syslogd on the named host.
• A comma separated list of users. Selected messages are written to those users if they are logged in.
• An asterisk. Selected messages are written to all logged-in users.

  Blank lines and lines whose first non-blank character is a hash (‘#’) character are ignored.

A configuration file might appear as follows:

# Log all kernel messages, authentication messages of
# level notice or higher and anything of level err or
# higher to the console.
# Don't log private authentication messages!
*.err;kern.*;auth.notice;authpriv.none  /dev/console

# Log anything (except mail) of level info or higher.
# Don't log private authentication messages!
*.info;mail.none;authpriv.none          /var/log/messages

# The authpriv file has restricted access.
authpriv.*                              /var/log/secure

# Log all the mail messages in one place.
mail.*                                  /var/log/maillog

# Everybody gets emergency messages, plus log them on another
# machine.
*.emerg                                 *
*.emerg                                 @arpa.berkeley.edu

# Root and Eric get alert and higher messages.
*.alert                                 root,eric

# Simplify security auditing, by collecting sudo uses.
! sudo
*.info                                  /var/log/sudo

# Collect time server reports.
#! ntpd
*.*                                     /var/log/ntpd

# Stop selecting on message tags.
!*

# Save mail and news errors of level err and higher in a
# special file.
uucp,news.crit                          /var/log/spoolerr

The effects of multiple selectors are sometimes not intuitive. For example ‘mail.crit,*.err’ will select the ‘mail’ facility messages at the level of ‘err’ or higher, not at the level of ‘crit’ or higher.

21.1 Standards ¶

The FTP server currently supports the following FTP requests. The letter case of any request is ignored.

The following non-standard, or UNIX specific, commands are supported by the SITE request.

The remaining FTP requests specified in RFC 959 are recognized, but not implemented. The extensions MDTM, REST, and SIZE are specified in RFC 3659, while EPRT and EPSV appear in RFC 2428, LPRT and LPSV in RFC 1639.

The ftp server will abort an active file transfer only when the ABOR command is preceded by a Telnet ‘Interrupt Process’ (IP) signal and a Telnet ‘Synch’ signal in the command Telnet stream, as described in Internet RFC 959. If a STAT command is received during a data transfer, preceded by a Telnet IP and Synch, transfer status will be returned.

ftpd interprets file names according to the globbing conventions used by csh. This allows users to utilize the metacharacters ‘*?[]{}~’.

The server applies the suggestions in RFC 2577, but the legacy behaviour with informational content in denials can be restored using the option --non-rfc2577.

21.2 Authentication ¶

ftpd authenticates users according to four rules.

1. The login name must be in the password data base, /etc/passwd, and must not have a null password. In this case a password must be provided by the client before any file operations can be performed.
2. The login name must not appear in the file /etc/ftpusers.
3. The user must have a standard shell. If the file /etc/shells exists and is readable, only programs listed there are considered standard shells. ftpd uses the C library function getusershell to enumerate standard shells. Standard shells are also known as valid login shells, valid user shells, or permitted user shells.
4. If the user name is ‘anonymous’ or ‘ftp’, an anonymous ftp account must be present in the password file (user ‘ftp’). In this case the user is allowed to log in by specifying any password (by convention an email address for the user should be used as the password).

A further access mechanism is provided by the file /etc/ftpchroot. A user mentioned therein will have all access confined to the subtree rooted at the home directory specified in /etc/passwd.

In the case of anonymous access, ftpd takes special measures to restrict the client’s access privileges. The server always performs a chroot to the home directory of the ‘ftp’ user.

In order that system security is not breached, it is recommended that the ‘ftp’ subtree be constructed with care, following these rules:

~ftp

Make the home directory owned by ‘root’ and not writable by anyone.

~ftp/bin

Make this directory owned by ‘root’ and not writable by anyone (mode 555). The program ls must be present to support the list command, unless the server was compiled with libls support. This program should be mode 111.

~ftp/etc

Make this directory owned by ‘root’ and not writable by anyone (mode 555). The files passwd and group must be present for the ls command to be able to produce owner names rather than numbers. The password field in passwd is not used, and should not contain real passwords. The file motd, if present, will be printed after a successful login. These files should be mode 444.

~ftp/pub

Make this directory mode 777 and owned by ‘ftp’. Guests can then place files which are to be accessible via the anonymous account in this directory.

21.3 Configuration files ¶

‘/etc/ftpchroot’

List of users to enclose in a chrooted directory. The anonymous user ‘ftp’ is always considered to be a member of this list, explicit or not.

‘/etc/ftpusers’

List of unwelcome/restricted users, always to be denied access.

‘/etc/ftpwelcome’

Welcome notice printed before server identification and any authentication exchange.

‘/etc/motd’

Welcome notice presented after completed login.

‘/etc/nologin’

If present, the contents are displayed and all further access is refused.

21.4 File format of ftpusers and ftpchroot. ¶

The files /etc/ftpusers and /etc/ftpchroot share a common file format. For better conformity with other implementations, each line is understood as consisting of fields separated by spaces, or by horizontal tabulators. Only the first non-empty field is examined at present. Both files are used for matching against a user name, desiring to use the FTP service.

Whenever the first printable character is a hash ‘#’, the input line is taken as a comment, and is ignored. Lines lacking non-empty fields are likewise ignored.

A field consisting of a single at-sign ‘@’, is treated as a wildcard and matches every input.

A field commencing with an at-sign ‘@’ and then continuing with an identifier, is understood as giving the name of a group. Should this name exist in /etc/groups, and the user name be a member of this same group, then the user name matches.

In all other cases, the field is taken as the identifier of a user, with which the requesting user is compared for verbatim match.

It is worthwhile to observe from the above cases, that a single ‘@’ on a line by itself in /etc/ftpchroot, will enforce chrooting upon every user allowed to access the FTP service. This gives a Draconian, protective configuration.

22.1 Invoking ¶

The only option is as follows:

-l ¶

--logging

Raise logging level for this service; use more than once for increased verbosity. The syslog facility in use is ‘LOG_DAEMON’.

Should rexecd have been built with PAM support, it reads any setting specified for a service named ‘rexec’.

22.2 Diagnostics ¶

Except for the last one listed below, all diagnostic messages are returned on the initial socket, after which any network connections are closed. An error is indicated by a leading byte with a value of 1 (0 is returned in step 7 above upon successful completion of all the steps prior to the command execution).

‘username too long’

The name is longer than 16 characters.

‘password too long’

The password is longer than 16 characters.

‘command too long’

The command line passed exceeds the size of the argument list (as configured into the system).

‘Login incorrect.’

No password file entry for the user name existed.

‘Password incorrect.’

The wrong password was supplied.

‘No remote directory.’

The chdir command to the home directory failed.

‘Try again.’

A fork by the server failed.

‘<shellname>: …’

The user’s login shell could not be started. This message is returned on the connection associated with the stderr, and is not ...

Note, that indicating ‘Login incorrect’ as opposed to ‘Password incorrect’ is a security breach which allows people to probe a system for users with null passwords.

23.1 Invoking ¶

The available options are as follows:

-4 ¶

--ipv4

Accept only IPv4 connections in daemon mode.

-6 ¶

--ipv6

Only IPv6 connections in daemon mode.

-a ¶

--verify-hostname

Ask hostname for verification.

-d[max] ¶

--daemon[=max]

Run in background daemon mode, optionally setting the maximal number of simultaneously running client sessions. The default limit is 10.

-D[level] ¶

--debug[=level]

Set debug level, not implemented.

-l ¶

--no-rhosts

Ignore client’s .rhosts file.

-L name ¶

--local-domain=name

Set local domain name, to which the server host belongs. By default the domain is recovered from the canonical name of the host.

-n ¶

--no-keepalive

Do not set SO_KEEPALIVE on sockets. This decreases the ability to close lost connections to once active clients.

-o ¶

--allow-root

Allow the root user to login, which is disallowed by default.

-p port ¶

--port=port

Listen on given port. Applicable only in daemon mode.

-r ¶

--reverse-required

Require reverse resolvability of remote host’s numerical IP.

For sites requiring improved authentication, Kerberos authentication is a viable decision, and possibly even with encryption for enhanced integrity. Three additional options are available for an executable rlogind compiled with Kerberos support.

-k ¶

--kerberos

Activate Kerberos authentication on all incoming requests.

-S name ¶

--server-principal=name

Set Kerberos server name, overriding canonical hostname.

-x ¶

--encrypt

Activate encryption of all data passed via the rlogind session. This may impact response time and CPU utilization, but provides increased security. Only for Kerberised mode of operation.

Should rlogind have been built with PAM support, it reads any setting specified for a service named either ‘rlogin’ or ‘krlogin’, the latter name for clients using Kerberised authentication.

23.2 Kerberos specific details ¶

The option -k is mandatory for Kerberised operation mode, while addition of the option -x will also demand encryption of every request to this particular server.

rlogind will, in Kerberised operation mode, as default instantiate itself using the principal name ‘host/canonical_name@DEFAULT_REALM’, a compound arranged from the running host’s canonical name, and from the default realm configured for the system. Either of these can be overridden using the option --server-principal, as follows:

rlogind -k -S alias.server.our
rlogind --kerberos --server-principal=@NEW.REALM
rlogind -k -x -S rlogin/backup.ex.org@OUR.REALM

When overriding only the realm, with the option -S, an initial at-sign is mandatory.

23.3 Protocol details ¶

When a service request is received, in non-Kerberised mode, the following protocol is initiated:

1. The server checks the client’s source port. If the port is not in the range 512-1023, the server aborts the connection.
2. The server next checks the client’s source address and requests the corresponding host name. If the hostname cannot be determined, the numerical representation of the host address is used. If the hostname is in the same domain as the server (according to the last two components of the domain name), or if the option -a is in effect, the address for the hostname is requested, verifying that the name and address correspond. Normal authentication is considered as failed, should this address verification fail.

Once the source port and address have been checked, rlogind proceeds with the authentication process as described in rshd: Remote shell server. The server then allocates a pseudo terminal, and manipulates file descriptors so that the slave half of the pseudo terminal becomes the stdin, stdout, and stderr for a login process. The login process is an instance of the login program, invoked with the option -f if authentication had succeeded. If automatic authentication had failed, the user is prompted to log in as if on a standard terminal line.

The parent of the login process manipulates the master side of the pseudo terminal, operating as an intermediary between the login process and the client instance of the rlogin program. In normal operation, the packet protocol described in ‘PTY’ is invoked to provide flow control using C-S/C-Q, and to propagate interrupt signals to the remote program. The login process transmits the client terminal’s baud rate, and its terminal type, as found in the environment variable TERM. The screen or window size of the terminal is requested from the client, and any later window size changes at the client’s side are propagated to the pseudo terminal as well.

Transport-level keepalive messages are enabled unless the option -n was in effect when starting rlogind. The use of keepalive messages allows sessions to be timed out, should the client crash, or otherwise become unreachable.

See ruserok in The GNU C Library Reference Manual, for details.

23.4 Diagnostics ¶

The exchange protocol states that a negotiation reaches a successful completion as soon as the server rlogind transmits back to the client a single null byte, marking the completion of all information exchange.

Error conditions are instead transmitted back to the client as a message containing an initial byte value 1, followed by a C-string indicating the cause of failure. All network connections are closed at the server side after this message. Some common messages follow:

‘Permission denied.’

The client presented insufficient credentials, or the client’s address is not sufficiently resolvable to pass the checks induced by options -a or -r.

‘Try again.’

A fork by the server failed.

24.1 Invoking ¶

The options are as follows:

-a ¶

--verify-hostname

Ask hostname for verification.

-k ¶

--kerberos

Use Kerberos authentication.

-l ¶

--no-rhosts

Ignore .rhosts file.

-L ¶

--log-sessions

Log successful logins.

-n ¶

--no-keepalive

Do not set SO_KEEPALIVE.

-S name ¶

--servername=name

Set Kerberos server name, overriding canonical hostname.

-v ¶

--vacuous

Fail any call asking for non-Kerberos authentication.

-r ¶

--reverse-required

Demand that the client’s IP address be resolvable as a host name.

Should rshd have been built with PAM support, it reads any setting specified for a service named either ‘rsh’ or ‘krsh’, the latter name for clients seeking Kerberised authentication.

24.2 Diagnostics ¶

Except for the last one listed below, all diagnostic messages are returned on the initial socket, after which any network connections are closed. An error is indicated by a leading byte with a value of 1 (0 is returned in step 10 above upon successful completion of all the steps prior to the execution of the login shell).

‘Locuser too long’

The name of the user on the client’s machine is longer than 16 characters.

‘Ruser too long’

The name of the user on the remote machine is longer than 16 characters.

‘Command too long’

The command line passed exceeds the size of the argument list (as configured into the system).

‘Login incorrect’

No password file entry for the user name existed.

‘Remote directory’

The chdir command to the home directory failed.

‘Permission denied’

The authentication procedure described above failed, or address resolution was insufficient.

‘Can't make pipe.’

The pipe needed for the stderr, wasn’t created.

‘Can't fork; try again.’

A fork by the server failed.

‘<shellname>: …’

The user’s login shell could not be started. This message is returned on the connection associated with the stderr, and is not preceded by a flag byte.

The authentication procedure used here assumes the integrity of each client machine and the connecting medium. This is insecure, but is useful in an “open” environment.

25.1 Invoking ¶

The following switches and options are available.

-a file ¶

--acl=file

Read site-wide ACLs from file.

-d ¶

--debug

Enable debugging.

-i seconds ¶

--idle-timeout=seconds

Set idle timeout length

-l ¶

--logging

Enable a somewhat enhanced logging verbosity, reporting attempted and dropped connections, as well as some more unexpected events that might arise.

-r seconds ¶

--request-ttl=seconds

Set time-to-live length for requests.

-S ¶

--strict-policy

Apply strict ACL policy on this system. This means that the site-wide ACL must provide explicit ‘allow’ rules for admitting traffic at all.

-t seconds ¶

--timeout=seconds

Set timeout length.

25.2 Modus operandi ¶

In normal operation, a client, the caller, initiates a rendezvous by sending a CTL_MSG of type ‘LOOK_UP’ to the server (see protocols/talkd.h). This causes the server to search its invitation tables to check whether an invitation currently exists for the caller (wanting to talk to the callee specified in the message). If the lookup fails, the caller then sends an ‘ANNOUNCE’ message causing the server to broadcast an announcement on the callee’s login ports requesting contact. When the callee responds, the local server uses the recorded invitation to respond with the appropriate rendezvous address and the caller and callee client programs establish a stream connection through which the conversation takes place.

This implementation offers an additional mechanism, whereby a site-wide access control list can be used to limit service access in general. For any local user, i.e., present on the server’s system, a further user owned file .talkrc is parsed, if at all present, in order to even further fine tune access to this particular user.

25.3 Access control in talkd ¶

The server can be run in a mode with additional access control, beyond the legacy capabilities of ntalkd. This is activated using the option -a, or equivalently --acl.

The format of this access control list is shared with the user specific file .talkrc. Normally the site-wide setting operates with a default value ‘allow’, but specifying the option -S, or --strict-policy, changes this default action to ‘deny’. In addition, the strict policy disables the possibility that an allowing action from the user specific ACL be able to override a denial resulting from the system-wide ACL setting.

As is usual, indentation, empty lines, and lines whose first printable character is the hash character, are all ignored. The general line format is

action user-exp [net-exp ...]

Each active line must contain at least two fields: an action and a user-exp.

The first field, action, must be either of ‘allow’ and ‘deny’. Any other value will lead to the line being ignored, but reported in the system log. Of course, the two values represent admitting and rejecting interpretations for the resulting rule.

The second field, user-exp, is a POSIX regular expression crafted to match user names. Remember that the regular expression would need anchors in order to test not only substrings.

It is important to note that in a site-wide ACL, the file selected by the switch -a, the expression user-exp is matched against the requested local user name, that of the callee.

While checking the callee’s private ACL-file .talkrc, the matching of user-exp is done against the remote caller’s name. Any other interpretation is plainly futile.

Each line may be augmented by a net list, containing one or more expressions net-exp. Each of these is either the simple word ‘any’, a numeric IPv4 address, or a full IPv4 address with an appended netmask. The effect is to restrict the applicability of the rule to the specified address range, or to set an explicit wildcard match ‘any’. The absence of a net list is equivalent to specifying a single ‘any’. The netmask can be specified as a CIDR mask length, or as an explicit address mask.

The actual evaluation is made separately for the site-wide ACL, and for the requested local user ACL, contained in the callee’s private file .talkrc. This latter file must be a regular file and must be owned by the very same user, have his primary group ownership, and not be group or world writeable. Should any of these prerequisites be violated, the user’s ACL is replaced by a single deny-all rule.

All rules in each set are evaluated, in the sense that whenever an expression net-exp matches the incoming IPv4 address, then the regular expression user-exp is tested for a match. That being the case, the corresponding action is recorded. The last match in each set determines the outcome in its category.

In the most common case, a system wide ‘deny’ is overridden if the local user has specified at least one valid and applicable rule, admitting access. In the contrary case, where no admitting user rule could be established at all, then a resulting ‘deny’, from a system wide ACL, will be used as the final action.

In strict policy mode, a site-wide ‘deny’ is always final, ignoring any user’s desire. The administrator must explicitly arrange some admitting rule, with an action ‘allow’, and some suitable net list. Still, the individual user can arrange his private file for an even narrower selection of friends.

26.1 Crafting an execution string. ¶

The server telnetd contains a built-in execution string which invokes login with arguments suitable for the operating system at hand. This preset choice corresponds to the standard use case of the service. For specialized purposes this implementation also offers a command line option -E, or --exec-login, to override the built-in execution of login, thus allowing almost any choice of handler.

A custom execution string could look like

telnetd -h -E '/usr/local/sbin/avrop  %t %U'

The execution string must as its first part provide an absolute path to an executable file. After that may follow arbitrary additional arguments. For this latter part, telnetd offers some replacement tokens that dynamically are replaced by content. All are of the form %<var>, where ‘<var>’ is a single letter from the following collection of selectors. A valid letter is called variable. The mark conditional, appearing below, indicates that the corresponding variable is conditionally assigned a value.

%a ¶

Returns ‘ok’ whenever authentication is complete. conditional

%d ¶

Produces a time and date string.

%h ¶

Gives the remote host name in canonical form.

%l ¶

States the local host name, also in canonical form.

%L ¶

Returns the path of the pseudo terminal assigned to the client.

%t ¶

Gives the terminal device stripped of the leading ‘/dev/’.

%T ¶

States the terminal type, like ‘xterm’. conditional

%u ¶

Provides the authenticated user name. conditional

%U ¶

Returns the user name passed as an environment variable USER by the remote client software. The value is empty, should the environment not provide a value.

In addition, a conditional construct is able to take one action in case a variable has an assigned value, and optionally to take another action in the opposite case. The construct is

%?<var>{true-stmt}[{false-stmt}]

The braces are here mandatory, while the brackets enclose the optional else-clause and are not included in actual use. The initial, motivating example, could thus be expanded to read

telnetd -h -E '/usr/local/sbin/avrop  %t %?a{%u krb5}{%U}'

In case authentication was completed as user ‘sigge’, the execution string would resolve to

/usr/local/sbin/avrop  pts/1 sigge krb5

In all other cases the result would be

/usr/local/sbin/avrop  pts/1 $USER

where $USER is the value of the corresponding environment variable and could possibly be empty.

tftpd [options] [directory ...]

27.1 Directory prefixes ¶

In addition to options, an invocation of tftpd can specify an optional list of directory prefixes. These are approved of according to two principles:

• Relative pathnames are ignored.
• At most twenty prefixes are approved, the rest is discarded.

A request for a file is decided upon as a consequence of evaluating these criteria:

• Every file request containing the substring ‘/../’ is denied, as is a file name beginning with ‘../’.
• Write requests must specify absolute locations.
• A file request, if specified as an absolute pathname, must begin with one of the approved directory prefixes, should at least one such prefix have been accepted.
• In the absence of a prefix collection, any absolute pathname is accepted, should the corresponding file exist.
• A file request, if specified as a relative name, will only be searched for below the acceptable prefixes, should at least one such prefix have been approved.
• A request for a relatively named file, is denied in the absence of approved directory prefixes.
• The resulting file must be world readable, or world writable, for a read request, or a write request, to succeed.

27.2 Use cases ¶

The standard use case is an entry in /etc/inetd.conf like

tftp dgram udp4 wait root /usr/sbin/tftpd \
       tftpd /tftpboot /altboot

This would allow the TFTP client to use any of

get kernel
get /tftpboot/kernel
get kernel.alt
get /altboot/kernel.alt
get /etc/motd

given that /tftpboot/kernel and /altboot/kernel.alt exist. Observe that also /etc/motd is accessible, in spite of there being no explicit mention of standard file locations.

A stronger mode of running a TFTP server is to use the ‘secure mode’, meaning that the serving process is running in a chrooted mode. Then a suitable configuration could be

tftp dgram udp4 wait root /usr/sbin/tftpd \
       tftpd --secure-dir=/srv/tftp-root  /tftpboot /altboot

Supposing the files kernel and kernel.alt to exist in the common directory /srv/tftp-root/altboot/, all the previously suggested client requests for a kernel would still be granted, but now any request for /etc/motd would be declined, and would get a reply ‘File not found’ back.

The chrooted setting is denying access outside of /srv/tftp-root, yet is not indicating this lock-in to the client, and is thus improving server isolation. Since neither of -u and -g were specified, the configuration reproduced above will in fact have the transmitting server process running with the default owner set to ‘nobody:nogroup’.

28.1 Options ¶

There is a single, specific option available:

-u location ¶

--uucico=location

Replace the hard coded location of uucico with the value specified as location.

28.2 Authentication steps. ¶

Invocation is expected to be conducted by a protocol described exchange of user name and password; unfortunately in clear text. If those agree with existing local entries, then uucpd verifies that the stated user also has user shell location identical to the full file system location of uucico. Should that not be the case, the request is declined.

For this latter check, the option --uucico is useful when setting the configuration for inetd. It is recommended to wrap the invocation line of uucpd within a call to tcpd in the standard fashion.

Copyright © 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc.
https://fsf.org/

Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.

  Copyright (C)  year  your name.
  Permission is granted to copy, distribute and/or modify this document
  under the terms of the GNU Free Documentation License, Version 1.3
  or any later version published by the Free Software Foundation;
  with no Invariant Sections, no Front-Cover Texts, and no Back-Cover
  Texts.  A copy of the license is included in the section entitled ``GNU
  Free Documentation License''.

    with the Invariant Sections being list their titles, with
    the Front-Cover Texts being list, and with the Back-Cover Texts
    being list.