doc/inetutils.texi - master - niceSOFT/inetutils

doc/inetutils.texi
5 346 строк172 KB
Collin Funk
telnetd: remove the --debug (-D) option
29 апр 2026, 09:46
29 апр 2026, 09:46aba8a39
\input texinfo @c -*-texinfo-*-
@c %**start of header
@setfilename inetutils.info
@settitle @sc{gnu} Inetutils
@c %**end of header

@include version.texi

@c Define new indices for file names and options.
@defcodeindex op
@defcodeindex fl

@c Put everything in one index (arbitrarily chosen to be the concept
@c index).
@syncodeindex fl cp
@syncodeindex fn cp
@syncodeindex ky cp
@syncodeindex op cp
@syncodeindex pg cp
@syncodeindex vr cp

@dircategory Basics
@direntry
* Inetutils: (inetutils).       GNU networking utilities.
* Common options: (inetutils)Common options.      Common options.
@end direntry

@dircategory Individual utilities
@direntry
* dnsdomainname: (inetutils)dnsdomainname invocation.       Show DNS domain name.
* ftp: (inetutils)ftp invocation.                 FTP client.
* ftpd: (inetutils)ftpd invocation.               FTP Daemon.
* hostname: (inetutils)hostname invocation.       Show or set system host name.
* ifconfig: (inetutils)ifconfig invocation.       Configure network interfaces.
* inetd: (inetutils)inetd invocation.             Internet super-server.
* logger: (inetutils)logger invocation.           Send messages to the system log.
* ping6: (inetutils)ping6 invocation.             Packets to IPv6 network hosts.
* ping: (inetutils)ping invocation.               Packets to network hosts.
* rcp: (inetutils)rcp invocation.                 Remote copy
* rexec: (inetutils)rexec invocation.             Remote execution client.
* rexecd: (inetutils)rexecd invocation.           Remote execution server.
* rlogin: (inetutils)rlogin invocation.           Remote login.
* rlogind: (inetutils)rlogind invocation.         Remote login server.
* rsh: (inetutils)rsh invocation.                 Remote shell.
* rshd: (inetutils)rshd invocation.               Remote shell server.
* syslogd: (inetutils)syslogd invocation.         Syslog server.
* talk: (inetutils)talk invocation.               Talk client.
* talkd: (inetutils)talkd invocation.             Talk server.
* telnet: (inetutils)telnet invocation.           User interface to TELNET.
* telnetd: (inetutils)telnetd invocation.         Telnet server.
* tftp: (inetutils)tftp invocation.               TFTP client.
* tftpd: (inetutils)tftpd invocation.             TFTP server.
* traceroute: (inetutils)traceroute invocation.   Trace the route to a host.
* uucpd: (inetutils)uucpd invocation.             Unix to Unix Copy.
* whois: (inetutils)whois invocation.             Whois user interface.
@end direntry

@copying
This manual documents version @value{VERSION} of the @sc{gnu} networking
utilities.

Copyright @copyright{} 2000--2026 Free Software Foundation, Inc.

@quotation
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''.
@end quotation
@end copying

@titlepage
@title @sc{gnu} @code{inetutils}
@subtitle GNU networking utilities
@subtitle for version @value{VERSION}, @value{UPDATED}
@author Alain Magloire et al.

@page
@vskip 0pt plus 1filll
@insertcopying
@end titlepage
@shortcontents
@contents

@ifnottex
@node Top
@top GNU Inetutils

@insertcopying
@end ifnottex

@menu
* Introduction::                       Caveats, overview, and authors.
* Common options::                     Common options.

Diagnostic programs

* dnsdomainname invocation::           Show DNS domain name.
* hostname invocation::                Show or set system host name.
* ifconfig invocation::                Configure network interfaces.
* logger invocation::                  Send messages to system log.
* ping invocation::                    Packets to network hosts.
* ping6 invocation::                   Packets to IPv6 network hosts.
* traceroute invocation::              Trace the route to a host.
* whois invocation::                   Whois user interface.

Clients

* ftp invocation::                     FTP client.
* rcp invocation::                     Remote copy
* rexec invocation::                   Remote execution client.
* rlogin invocation::                  Remote login.
* rsh invocation::                     Remote shell.
* talk invocation::                    Talk client.
* telnet invocation::                  User interface to TELNET.
* tftp invocation::                    TFTP client.

Daemons

* inetd invocation::		       Internet super-server.
* syslogd invocation::		       Syslog server.
* ftpd invocation::		       FTP Daemon.
* rexecd invocation::		       Remote execution server.
* rlogind invocation::		       Remote login server.
* rshd invocation::		       Remote shell server.
* talkd invocation::		       Talk server.
* telnetd invocation::		       Telnet server.
* tftpd invocation::		       TFTP server.
* uucpd invocation::		       Unix to Unix Copy.

Appendix

* GNU Free Documentation License::     The license for this manual.
* Index::                              Index of manual.
@end menu

@c OK -- 2009-04-27
@node Introduction
@chapter Introduction
@cindex introduction

The GNU Networking Utilities is a distribution of common networking
utilities and servers, including for example ping, traceroute and ftp.

This manual is a work in progress: many sections make no attempt to
explain basic concepts in a way suitable for novices.  Thus, if you
are interested, please get involved in improving this manual.  The
entire @sc{gnu} community will benefit.

@cindex bug, reporting
Please report bugs to @email{bug-inetutils@@gnu.org}.  Remember to
include the version number, machine architecture, input files, and any
other information needed to reproduce the bug: your input, what you
expected, what you got, and why it is wrong.  Diffs are welcome, but
please include a description of the problem as well, since this is
sometimes difficult to infer.

The individual utilities were originally derived from the 4.4BSDLite2
distribution, although some of them have more or less been rewritten.
What you are reading now is the authoritative and complete
documentation for these utilities; the man pages are now automatically
generated.

Many features were integrated from NetBSD, OpenBSD, FreeBSD and
GNU/Linux, the merges were done by a group of dedicated hackers (in no
particular order): Jeff Bailey, Marcus Brinkmann, Michael Vogt,
Bernhard Rosenkraenzer, Kaveh R. Ghazi, NIIBE Yutaka, Nathan
Neulinger, Jeff Smith, Dan Stromberg, David O'Shea, Frederic Goudal,
Gerald Combs, Joachim Gabler, Marco D'Itri, Sergey Poznyakoff, and
many more.

@node Common options
@chapter Common options
@cindex common options

Certain options are available in all these programs.  Rather than
writing identical descriptions for each of the programs, they are
described here.  (In fact, every @sc{gnu} program accepts, or should
accept, these options.)

Many of these programs take arbitrary strings as arguments.  In those
cases, @option{--help} and @option{--version} are taken as these
options only if there is one and exactly one command line argument.

@table @option
@item --help
@opindex --help
@cindex help, online
Print a usage message, listing all available options, then exit
successfully.

@item --usage
@opindex --usage
@cindex usage, online
Print a condensed usage message, displaying all available options
formatted like a command line call, then exit successfully.

@item --version
@opindex --version
@cindex version number, finding
Print the version number, then exit successfully.

@item --
@opindex --
@cindex option delimiter
Delimit the option list.  Later arguments, if any, are treated as
operands even if they begin with @samp{-}.
@end table

@menu
* Exit status::                 Indicating program success or failure.
@end menu

@node Exit status
@section Exit status

@macro exitstatus
An exit status of zero indicates success, and a nonzero value
indicates failure.
@end macro

Nearly every command invocation yields an integral @dfn{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 @samp{1}, though
it may differ on unusual platforms, as POSIX requires only that it be
nonzero.

@node dnsdomainname invocation
@chapter @command{dnsdomainname}: Show DNS domain name
@pindex dnsdomainname

@command{dnsdomainname} is a program to show the domain part of the
system's fully qualified domain name.  For example, if the FQDN of the
system is @code{name.example.org} the command will show
@code{example.org}.
The output is not necessarily related to the NIS/YP domain name.

The tool uses gethostname to get the host name of the system and
then getaddrinfo to resolve it into a canonical name.
The domain part of the canonical name is shown, i.e., the part
after the first period@tie{}(@code{.}) of the official name.

@noindent
Synopsis:

@example
dnsdomainname [@var{option}@dots{}]
@end example

@noindent
There is no command specific option.

@node hostname invocation
@chapter @command{hostname}: Show or set system host name.
@pindex hostname

@command{hostname} is a program to show or to set the name of a
host system.

@noindent
Synopsis:

@example
hostname [@var{option}@dots{}]
hostname @var{name}
@end example

@noindent
where @var{name} is the name to be used by the running host.

@section Command line options
@anchor{hostname options}

@table @option
@item -a
@itemx --aliases
@opindex -a
@opindex --aliases
Get alias names.

@item -d
@itemx --domain
@opindex -d
@opindex --domain
Get DNS domain name.

@item -f
@itemx --fqdn
@itemx --long
@opindex -f
@opindex --fqdn
@opindex --long
Get DNS host name or Fully Qualified Domain Name.

@item -F @var{file}
@itemx --file=@var{file}
@opindex -F
@opindex --file
Set host name or NIS domain name from FILE.

@item -i
@itemx --ip-addresses
@opindex -i
@opindex --ip-addresses
Get addresses for the host name.

@item -s
@itemx --short
@opindex -s
@opindex --short
Get short host name.

@item -y
@itemx --yp
@itemx --nis
@opindex -y
@opindex --yp
@opindex --nis
Get NIS/YP domain name.
@end table

@node ifconfig invocation
@chapter @command{ifconfig}: Configure network interfaces
@pindex ifconfig

@command{ifconfig} is a program to retrieve and to set selected
properties of network interfaces.  It is best viewed as a tool to
get information, rather than for changing the behaviour of adapters,
since it is hard to support property setting in a portable manner.

@noindent
Synopsis:

@example
ifconfig @var{iface} [@var{arg}@dots{}]
ifconfig -i @var{iface} [@var{option}@dots{}] [-i @var{iface2} [@var{option}@dots{}]]
@end example

@section Command line options
@anchor{ifconfig options}

@table @option
@item -a
@itemx --all
@opindex -a
@opindex --all
Display all available interfaces, including those that not are
marked as `up', i.e., also the inactive interfaces.

@item -A @var{addr}
@itemx --address=@var{addr}
@opindex -A
@opindex --address
Set address of selected interface to @var{addr}.

@item -b @var{addr}
@itemx -B @var{addr}
@itemx --brdaddr=@var{addr}
@itemx --broadcast=@var{addr}
@opindex -b
@opindex -B
@opindex --brdaddr
@opindex --broadcast
Set broadcast address of selected interface to @var{addr}.

@item -d @var{addr}
@itemx -p @var{addr}
@itemx --dstaddr=@var{addr}
@itemx --peer=@var{addr}
@opindex -d
@opindex -p
@opindex --dstaddr
@opindex --peer
Set destination (peer) address of selected interface.

@item --down
@opindex --down
Deactivate the selected interface.

@item -F @var{list}
@itemx --flags=@var{list}
@opindex -F
@opindex --flags
Change those interface flags mentioned in @var{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 @samp{no}, the corresponding flag is cleared.
The output of @command{ifconfig} with the option @option{--help}
contains a list of available flag names.

@item --format=@var{format}
@opindex --format
Select output format; the value @samp{help} prints a list
of all available formats.

@item -i @var{name}
@itemx --interface=@var{name}
@opindex -i
@opindex --interface
Select the named interface for any following action.

@item -l
@itemx --list
@opindex -l
@opindex --list
List, with name only, all available interfaces, or only those
selected should at least one option @option{-i} have specified.

@item -m @var{mask}
@itemx --netmask=@var{mask}
@opindex -m
@opindex --netmask
Set netmask of selected interface to @var{mask}.

@item --metric=@var{n}
@opindex --metric
Set the metric of selected interface to the number @var{n}.

@item -M @var{n}
@itemx --mtu=@var{n}
@opindex -M
@opindex --mtu
Set MTU of selected interface to the number @var{n}.

@item -s
@itemx --short
@opindex -s
@opindex --short
Use short output format.  This is identical to specifying
@samp{--format=netstat}.

@item --up
@opindex --up
Activate the selected interface.

@item -v
@itemx --verbose
@opindex -v
@opindex --verbose
Print informational messages when configuring an interface.
@end table

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

@section Formatted status output
@anchor{ifconfig formats}

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
@option{--format=help}.  In the following table the valid formats
are given, each is used in the form @option{--format=@var{name}}.

@table @asis
@item check
@itemx check-existence
@itemx ?
Place holders for the ability to check whether the interfaces
selected by one or more options @option{-i} are determining
existing interfaces in the running system.  No output in case
of success, an error message in case of a failure.

@item gnu
@itemx default
Standard GNU output format.

@item gnu-one-entry
Like the previous format, but with intermediary newlines removed.

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

@item net-tools
Imitation of presentation used by the implementation in @samp{net-tools}.
Default format for GNU/Linux.

@item netstat
Terse output with statistics, similar to that of @code{netstat -i}.

@item osf
Format variant of @samp{unix} preferred by OSF's implementation.

@item unix
Traditional UNIX type format.  Default for BSD, HPUX and Solaris.
@end table

@section Legacy syntax
@anchor{ifconfig legacy syntax}

The traditional mode of invoking @command{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.

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

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
@samp{ifconfig --usage}.

@node logger invocation
@chapter @command{logger}: Send messages to system log
@pindex logger

@command{logger} is a program to send entries to system log.  It
provides a shell command interface similar to the system log module.
For background information,
@pxref{Syslog, , Syslog, libc, The GNU C Library Reference Manual}.

@noindent
Synopsis:

@example
logger [@var{option}@dots{}] [@var{message}]
@end example

@section Command line options
@anchor{logger options}

@table @option
@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Use IPv4 as transport when logging to a host.  The default behaviour
is to use whatever IP version that matches the host.

@item -6
@itemx --ipv6
@opindex -6
@opindex --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 @option{--ipv4} or @option{--ipv6} is stated.

@item -f @var{file}
@itemx --file=@var{file}
@opindex -f
@opindex --file
Log the content of the specified file.  If @var{file} is @samp{-} then
standard input is assumed.

@item -h @var{host}
@itemx --host=@var{host}
@opindex -h
@opindex --host
Send messages to the given host or socket.  The @var{host} argument
can be either a local UNIX socket name (containing a slash @samp{/}),
or be of the form

@example
@var{host}[:@var{port}]
@end example

@noindent
where @var{host} is the remote host name or IP address, and the
optional @var{port} is a decimal port number or symbolic service
name from @file{/etc/services}.  If @var{port} is not specified,
the port number corresponding to the @samp{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]).

@item -i[@var{pid}]
@itemx --id=[@var{pid}]
@opindex -i
@opindex --id
Add process ID to each message.  If @var{pid} is not supplied, use the
process ID of the logger process with each line.  Notice, that
@var{pid} is an optional argument.  When supplied to the @option{-i}
option, it must follow the @samp{i} letter immediately, without any
separating whitespace.  When supplied to the @option{--id} form, it
must be separated from it by exactly one equals sign.

@item -p @var{priority}
@itemx --priority=@var{priority}
@opindex -p
@opindex --priority
Enter the message with the specified priority.  The priority may be
specified numerically or as a @samp{facility.level} pair.  For
example, @option{-p local3.info} logs the message at the informational
level in the @samp{local3} facility.  The default is
@samp{user.notice}.

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

@item -s
@itemx --stderr
@opindex -s
@opindex --stderr
Log the message to standard error, as well as to the system log.

@item -S @var{addr}
@itemx --source=@var{addr}
@opindex -S
@opindex --source
Supply the source IP address for INET connections.  This option is
useful in conjunction with @option{--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.

@item -t @var{tag}
@itemx --tag=@var{tag}
@opindex -t
@opindex --tag
Mark every line in the log with the specified tag.

@item -u @var{socket}
@itemx --unix=@var{socket}
@opindex -h
@opindex --host
Send messages to the given local UNIX socket.  The @var{socket} argument
can be either an absolute path (starting with a slash @samp{/}), or a relative
path understood relative to the current working directory.
@end table

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

@section Examples
@anchor{logger examples}

The following examples illustrate the usage of the @command{logger}
command:

@enumerate 1
@item Log the message @samp{System rebooted} to the local syslog.
Use default facility and priority:

@example
logger System rebooted
@end example

@item Run command and send its error output to the channel
@samp{local0.err}.  Mark each message with tag @samp{cmd}:

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

@item Log each line from file @file{warnings} to channel
@samp{daemon.warn} on host @samp{logger.runasimi.org},
using the source IP @samp{10.10.10.1}:

@example
logger -p daemon.warn -h logger.runasimi.org -S 10.10.10.1  \
@verb{|       |}--file warnings
@end example
@end enumerate

@node ping invocation
@chapter @command{ping}: Packets to network hosts
@pindex ping

@c FIXME: The text is far to detailed about the actual implementation
@c of ping.  A user doesn't need to know that we are using TIMEVAL, or
@c how things are padded.

@command{ping} uses ICMP datagrams to provoke a response
from the chosen destination host, mainly intending to probe
whether it is alive.

The used datagram, of type @code{ECHO_REQUEST}, contains some header
information and some additional payload, usually a time stamp.
By a suitable choice of payload, different host or router properties
are detectable, as the emitted datagram travels to its destination.

@ignore %* Too detailed for end user.
@command{ping} uses the ICMP protocol's mandatory
@code{ECHO_REQUEST} datagram to elicit an ICMP type @code{ECHO_REPLY}
packet from a host or gateway.
@code{ECHO_REQUEST} datagrams (@dfn{pings}) have an IP and
an ICMP header, followed by a @dfn{struct timeval} and then
an arbitrary number of @dfn{padding} bytes used to fill out the packet.
@end ignore

@noindent
Synopsis:

@example
ping [@var{option}@dots{}] @var{host}
@end example

@noindent
Sending echo requests is the standard use of @command{ping},
but by far not the only use case.

@menu
* Ping options::
* Fault isolation::
* Duplicate and damaged packets::
* Data patterns::
* TTL details::
* Further remarks::
@end menu

@node Ping options
@section Command line options
@anchor{ping options}

@c Options controlling ICMP request types:
@c       --address              Send ICMP_ADDRESS packets (root only)
@c       --echo                 Send ICMP_ECHO packets (default)
@c       --mask                 Same as --address
@c       --timestamp            Send ICMP_TIMESTAMP packets
@c   -t, --type=TYPE            Send TYPE packets

Selection of packet type is handled by these first options:

@table @option
@item --address
@opindex --address
Send ICMP_ADDRESS packets, thus requesting the address netmask
in use by the targeted host.

@item --echo
@opindex --echo
Send ICMP_ECHO requests.
This is the default action.

@item --mask
@opindex --mask
Identical to @option{--address}.

@item --timestamp
@opindex --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@tie{}UTC.
The first of these, @samp{icmp_otime}, contains the original
time of sending the request.
Then comes @samp{icmp_rtime}, the time of reception by the target,
and finally, @samp{icmp_ttime}, the time of transmitting an answer
back to the originator.

@item -t @var{type}
@itemx --type=@var{type}
@opindex --type
@opindex -t
Send @var{type} packets.  Accepted values are @samp{address},
@samp{echo}, @samp{mask}, and @samp{timestamp}.
@end table

@c Options valid for all request types:
@c   -c, --count=NUMBER         Stop after sending NUMBER packets
@c   -d, --debug                Set the SO_DEBUG option
@c   -i, --interval=NUMBER      Wait NUMBER seconds between sending each packet
@c   -n, --numeric              Do not resolve host addresses
@c   -r, --ignore-routing       Send directly to a host on an attached network
@c   -T, --tos=NUM              Set type-of-service to NUM
@c       --ttl=NUMBER           Set specified time-to-live on packet

@noindent
The following options are available for all packet types:

@table @option
@item -c @var{n}
@itemx --count=@var{n}
@opindex -c
@opindex --count
Stop after sending and receiving answers to a total of
@var{n} packets.

@item -d
@itemx --debug
@opindex -d
@opindex --debug
Set the SO_DEBUG option on the socket being used.

@item -i @var{n}
@itemx --interval=@var{n}
@opindex -i
@opindex --interval
Wait @var{n} seconds until sending next packet.
The default is to wait for one second between packets.
This option is incompatible with the option @option{-f}.

@item -n
@itemx --numeric
@opindex -n
@opindex --numeric
Numeric output only.  No attempt will be made to resolve
symbolic names for host addresses.

@item -r
@itemx --ignore-routing
@opindex -r
@opindex --ignore-routing-log
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 @command{routed}).

@item -T @var{num}
@itemx --tos=@var{num}
@opindex -T
@opindex --tos
Set type-of-service, TOS field, to @var{num} on
transmitted packets.

@item --ttl=@var{n}
@opindex --ttl
Set the specified number @var{n} as value of time-to-live when
transmitting packets.  Acceptable values are 1 to 255, inclusive.

@item -v
@itemx --verbose
@opindex -v
@opindex --verbose
Produce more verbose output, giving more statistics.

@item -w @var{n}
@itemx --timeout=@var{n}
@opindex -w
@opindex --timeout
Stop after @var{n} seconds.

@item -W @var{n}
@itemx --linger=@var{n}
@opindex -W
@opindex --linger
Maximum number of seconds @var{n} to wait for a response.
@end table

@c Options valid for --echo requests:
@c   -f, --flood                Flood ping (root only)
@c       --ip-timestamp=FLAG    Timestamp IP option of types tsonly,
@c                              tsaddr, or (not yet implemented) prespec.
@c   -l, --preload=NUMBER       Send NUMBER packets as fast as possible before
@c                              falling into normal mode of behavior (root only)
@c   -p, --pattern=PATTERN      Fill ICMP packet with given pattern (hex)
@c   -q, --quiet                no packet message
@c   -R, --route                Record route IP option
@c   -s, --size=NUMBER          Send NUMBER data octets

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.

@table @option
@item -f
@itemx --flood
@opindex -f
@opindex --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 @samp{.} 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.

@item --ip-timestamp=@var{flag}
@opindex --ip-timestamp
Include IP option Timestamp in transmitted packets.
The value @var{flag} is either @samp{tsonly}, which only records up
to nine time stamps, or @samp{tsaddr}, which records IP
addresses as well as time stamps, but for at most four hosts.

@item -l @var{n}
@itemx --preload=@var{n}
@opindex -l
@opindex --preload
If @var{n} is specified, ping sends that many packets as fast as
possible before falling into its normal mode of operation.

@item -p @var{pat}
@itemx --pattern=@var{pat}
@opindex -p
@opindex --pattern
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, @option{-p ff} will cause the sent packet to be filled
with all ones.

@item -q
@itemx --quiet
@opindex -q
@opindex --quiet
Do not print timing for each transmitted packet.
@item -R
@itemx --route
@opindex -R
@opindex --route
Record route.  Includes the @code{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.

@item -s @var{n}
@itemx --size=@var{n}
@opindex -s
@opindex --size
Specifies the number of data bytes to be sent.  The default is 56,
which translates into 64@tie{}ICMP data bytes, taking
the 8@tie{}bytes of ICMP header data into account.
@end table

@node Fault isolation
@section Using ping for network fault isolation

When using @command{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 @samp{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.

@ignore
@section ICMP Packet Details

An IP header without options consists of 20 bytes.
An ICMP type ECHO_REQUEST packet contains an additional
8 bytes worth of ICMP header followed by an arbitrary
amount of data.
When a packet size is stated, that indicates
the size of the extra piece of data (the default is 56).
Thus the amount of data received as an IP packet
with an ICMP type ECHO_REPLY, will always be 8 bytes larger
than the requested data space (the ICMP header).

If the data space is at least eight bytes large, ping uses the first
eight bytes of this space to include a timestamp which it uses in the
computation of round trip times.  If less than eight bytes of pad are
specified, no round trip times are given.
@end ignore

@node Duplicate and damaged packets
@section 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).

@node Data patterns
@section 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 @option{-p} option of ping.

@node TTL details
@section TTL details

The TTL field, @dfn{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 @code{ECHO_REQUEST})
packets to 255.  This is why you will find you can ping some hosts,
but not reach them with @command{telnet} or @command{ftp}.

During normal operation, @command{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:

@itemize @bullet
@item
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.

@item
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.

@item
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.

@end itemize

@node Further remarks
@section Further observations
Many hosts and gateways ignore the @code{RECORD_ROUTE} field, since
the maximum IP header length is far to small to hold all
the routes.
There is not much that can be done about this.

Flood pinging is not recommended in general, and flood pinging the
broadcast address should only be done under very controlled
conditions.

Some BSD variants offer a kernel setting to inhibit all replies
to ICMP_MASKREQ packets, but in general, Unices are designed either
to answer the request with a valid netmask, or to drop the request,
causing @command{ping} to wait for a timeout condition.

@node ping6 invocation
@chapter @command{ping6}: Packets to IPv6 network hosts
@pindex ping6

@command{ping6} uses ICMPv6 datagrams to get a response
from the chosen destination host.
The most common use is to probe whether the remote
system is responsive.
Observe that this program only uses IPv6 datagrams.

Each datagram, of type @code{ECHO_REQUEST}, carries some header
information and some additional payload, usually a time stamp.
Making a suitable choice of payload, it is possible to probe
different host or router properties on the way as the emitted
datagram travels to its destination.

@noindent
Synopsis:

@example
ping6 [@var{option}@dots{}] @var{host}
@end example

@noindent
Sending simple, timed echo requests is the standard use
of @command{ping6}, but is by far not the only use case.

This command is a close parallel to @command{ping},
except that it handles IPv6 and is thus not able
to handle peculiarities of IPv4.

@section Command line options
@anchor{ping6 options}

@table @option
@item -c @var{n}
@itemx --count=@var{n}
@opindex -c
@opindex --count
Stop after sending and receiving answers to a total of
@var{n} packets.

@item -d
@itemx --debug
@opindex -d
@opindex --debug
Set the SO_DEBUG option on the socket being used.

@item -f
@itemx --flood
@opindex -f
@opindex --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 @samp{.} 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!

@item --hoplimit=@var{n}
@opindex --hoplimit
Limit maximal distance to @var{n}.
Acceptable values are 1 to 255, inclusive.

@item -i @var{n}
@itemx --interval=@var{n}
@opindex -i
@opindex --interval
Wait @var{n} seconds until sending next packet.
The default is to wait for one second between packets.
This option is incompatible with the option @option{-f}.

@item -l @var{n}
@itemx --preload=@var{n}
@opindex -l
@opindex --preload
Sends @var{n} packets as fast as possible before falling
back to the normal mode of operation.

@item -n
@itemx --numeric
@opindex -n
@opindex --numeric
Numeric output only.
No attempt will be made to resolve symbolic names for host addresses.

@item -p @var{pattern}
@itemx --pattern=@var{pattern}
@opindex -p
@opindex --pattern
Up to 16 hexadecimal pad bytes are given as @var{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, @option{-p ff} will cause the sent packets
to have payloads with every bit set to one.

@item -q
@itemx --quiet
@opindex -q
@opindex --quiet
Do not print timing result of each transmitted packet.

@item -r
@itemx --ignore-routing
@opindex -r
@opindex --ignore-routing-log
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 @command{routed}.

@item -s @var{n}
@itemx --size=@var{n}
@opindex -s
@opindex --size
Specifies the number of data bytes to be sent.  The default is 56,
which translates into 64@tie{}ICMP data bytes, taking
the 8@tie{}bytes of ICMP header data into account.

@item -T @var{num}
@itemx --tos=@var{num}
@opindex -T
@opindex --tos
Set the traffic class to @var{num} on transmitted packets.

@item --ttl=@var{n}
@opindex --ttl
Synonym for @option{--hoplimit}.

@item -v
@itemx --verbose
@opindex -v
@opindex --verbose
Produce more verbose output, giving more statistics.

@item -w @var{n}
@itemx --timeout=@var{n}
@opindex -w
@opindex --timeout
Stop after @var{n} seconds.
@end table

The documentation of @command{ping} provides several
pieces of information, and discussions, relevant to
the use of @command{ping6}.
Keep in mind, though, that the differing address family
causes some discrepancy.
@xref{ping invocation}.

@node traceroute invocation
@chapter @command{traceroute}: Trace the route to a host
@pindex traceroute

@command{traceroute} prints a trace of the route
IP@tie{}packets are travelling to a remote host.

@noindent
Synopsis:

@example
traceroute [@var{option}@dots{}] @var{host}
@end example

@section Command line options
@anchor{traceroute options}

@table @option
@item -f @var{num}
@itemx --first-hop=@var{num}
@opindex -f
@opindex --first-hop
Set the initial hop distance to @var{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.

@item -g @var{gates}
@itemx --gateways=@var{gates}
@opindex -g
@opindex --gateways
Set intermediary hosts used in loose source routing.
The argument @var{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 @option{-g} produce a concatenated list.

@item -I
@itemx --icmp
@opindex -I
@opindex --icmp
Use ICMP ECHO datagrams for probing the remote host.

@item -m @var{num}
@itemx --max-hop=@var{num}
@opindex -m
@opindex --max-hop
Set the maximum time-to-live allowed for probing.
In other words, stop probing when the hop distance
is in excess of @var{num}.
The default limit is 64.

@item -M @var{method}
@itemx --type=@var{method}
@opindex -M
@opindex --type
Use @var{method} as carrier packets for traceroute operations.
Supported choices are @samp{icmp} and @samp{udp}, where @samp{udp}
is the default type.

@item -p @var{port}
@itemx --port=@var{port}
@opindex -p
@opindex --port
Set destination port of target to @var{port}.
The default value is 33434.

@item -q @var{num}
@itemx --tries=@var{num}
@opindex -q
@opindex --tries
Send a total of @var{num} probe packets per hop, defaulting to 3.

@item --resolve-hostnames
@opindex --resolve-hostnames
Attempt to resolve all addresses as hostnames.

@item -t @var{num}
@itemx --tos=@var{num}
@opindex -t
@opindex --tos
Set type-of-service, TOS field, to @var{num} on
transmitted packets.

@item -w @var{num}
@itemx --wait=@var{num}
@opindex -w
@opindex --wait
Set timeout in seconds, within which a returning response
packet is accepted as such.
Default waiting time is three seconds.
@end table

@section Diagnostic tokens
@anchor{traceroute printing}

During execution, @command{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 @samp{*} 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:

@table @samp
@item !F
Fragmentation needed by gateway.

@item !H
Host not reachable from gateway.

@item !N
Network not reachable from gateway.

@item !P
Protocol not usable at host, or within network.

@item !S
Source routing failed at gateway.

@item !T
Host or network not reachable for stated type of service, TOS.

@item !U
Isolated host, not reachable.

@item !X
Forbidden by remote administration.
@end table

@node whois invocation
@chapter @command{whois}: User interface to WHOIS data bases.
@pindex whois

The functionality of a world wide Internet is dependent on
stored node information of different kinds.
Registrars keep much relevant material in WHOIS data bases.
This utility @command{whois} is able to query those sources
for general and for particular properties of most domains.

For many domains there are names of suitable data base servers
hard coded into @command{whois}, ready to query for domain
relevant information.
Since servers' names do change from time to time,
this utility might occasionally need some guidance using
a suitable command line option.

@noindent
Synopsis:

@example
whois [@var{OPTION}@dots{}] @var{OBJECT}@dots{}
@end example

@section Command line options
@anchor{whois options}

@table @option
@item -a
@opindex -a
Search all data bases.

@item -F
@opindex -F
Fast and raw output. Implies @option{-r}.

@item -g @var{source}:@var{first}-@var{last}
@opindex -g
Find updates for an object from provider @var{source},
starting from the version with serial key @var{first},
and ending with serial key @var{last}.

@item -h @var{host}
@itemx --server=@var{host}
@opindex -h
@opindex --server
Connect to server @var{host}.

@item -H
@opindex -H
Hide legal disclaimers.

@item -i @var{attr}[,@var{attr2}@dots{}]
@opindex -i
Do an inverse lookup for specified attributes.
Use a comma separated list for multiple attributes.

@item -l
@opindex -l
One level less specific lookup.
Applies to RPSL only.

@item -L
@opindex -L
Find all less specific matches.

@item -m
@opindex -m
Find more specific matches, one level deeper.

@item -M
@opindex -M
Find all more specific matches.

@item -p @var{port}
@opindex -p
Connect to server port @var{port}.

@item -q @{version|sources@}
@opindex -q
Query specified server info.
Applies to RPSL only.

@item -r
@opindex -r
Turn off recursive lookups.

@item -R
@opindex -R
Force output to show local copy of the domain object,
even if it contains a referral.

@item -s @var{source}[,@var{source2}@dots{}]
@opindex -s
Search the data base at @var{source}.
A comma separated list queries multiple providers.

@item -S
@opindex -S
Tell server to refrain from syntactic sugar.

@item -t @var{type}
@opindex -t
Request a template for objects of type @var{type}.
Use the value @samp{all} for a list of possible types.

@item -T @var{type}[,@var{type2}@dots{}]
@opindex -T
Search only for objects of type @var{type}.
A comma separated list allows for multiple types.

@item -V
@itemx --verbose
@opindex -V
@opindex --verbose
Verbosely explain all actions taken.

@item -x
@opindex -x
Search only for exact matches.
Applicable only to RPSL.
@end table

@section Environment variables
@anchor{whois environment}

@command{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 @option{-h}.
If neither of these have a say, then the default server to ask
is @samp{whois.internic.net}, but this name is in turn overruled
by a server name in the environment variable @env{WHOIS_SERVER}.

@table @env
@item LANG
When the server @samp{whois.nic.ad.jp} is queried, and only then,
any non-Japanese locale in @env{LANG} will ask the server to reply
with English text, not Japanese.

@item WHOIS_HIDE
When set, the effect on @command{whois} is as if the
option @option{-H} had been given.

@item WHOIS_SERVER
Data base server to query when internal hinting is inconclusive.
When unset, @samp{whois.internic.net} is used as default server.

@end table

@node ftp invocation
@chapter @command{ftp}: FTP client
@pindex ftp

@command{ftp} is the user interface to FTP,
the File Transfer Protocol.
The program allows a user to transfer files to and from a remote
network site.

@noindent
Synopsis:

@example
ftp [@var{option}@dots{}] [@var{host} [@var{port}]]
pftp [@var{option}@dots{}] [@var{host} [@var{port}]]

ftp [@var{option}@dots{}] @var{user@@host} [@var{port}]
pftp [@var{option}@dots{}] @var{user@@host} [@var{port}]
@end example

@noindent
The alternate name @command{pftp} is starting in passive mode,
but is otherwise identical to @command{ftp}.

The client host with which @command{ftp} is to communicate may be
specified on the command line.
If this is done, @command{ftp} will immediately attempt to establish
a connection to the FTP server running on that host.
Optionally, a remote user name can be specified at will.
Otherwise, the program will start a command interpreter and will await
further instructions from the user.
Commands can either be entered interactively,
or piped as a batched job read from standard input.
@command{ftp} is able to distinguish between these two modes
of operation.

@menu
* Ftp options::
* Ftp commands::
* Ftp environment::
* Aborting a file transfer::
* File naming conventions::
* File transfer parameters::
* The .netrc file::
@end menu

@node Ftp options
@section Command line options
@anchor{ftp options}

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

@table @option
@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Initially set addressing to IPv4 only.

@item -6
@itemx --ipv6
@opindex -6
@opindex --ipv6
Initially set addressing to IPv6 only.

@item -A
@itemx --active
@opindex -A
@opindex --active
Enable active mode transfer.  Default mode for @command{ftp}.

@item -d
@itemx --debug
@opindex -d
@opindex --debug
Enable debugging output and possibly also socket debugging.

@item -e
@itemx --no-edit
@opindex -e
@opindex --no-edit
Disables the editing of commands.  This is default setting for batch
mode, without a TTY, or when the environment variable @env{TERM} is not
set or its value is @samp{dumb}.

@item -g
@itemx --no-glob
@opindex -g
@opindex --no-glob
Disables file name globbing.

@item -i
@itemx --no-prompt
@opindex -i
@opindex --no-prompt
Turns off interactive prompting during multiple file transfers.

@item -N @var{netrc}
@itemx --netrc=@var{netrc}
@opindex -N
@opindex --netrc
Set a preferred location of the @file{.netrc} file,
thus overriding any environment setting in @env{NETRC},
as well as the default location @file{$HOME/.netrc},
@pxref{The .netrc file}.

@item -n
@itemx --no-login
@opindex -n
@opindex --no-login
Restrains @command{ftp} from attempting @dfn{auto-login} upon initial
connection.  If auto-login is enabled, @command{ftp} will check the
@file{.netrc} (@pxref{The .netrc file}) file in the user's home
directory for an entry describing an account on the remote machine.
If no entry exists, @command{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.

@item -p
@itemx --passive
@opindex -p
@opindex --passive
Enable passive mode transfer.  Default mode when invoked
as @command{pftp}.

@item --prompt[=@var{prompt}]
@opindex --prompt
Print a command-line prompt, even if not on a tty.  If @var{prompt} is
supplied, its value is used instead of the default @samp{ftp> }.
Notice, that the argument is optional.

@item -t
@itemx --trace
@opindex -t
@opindex --trace
Enable packet tracing (not implemented).

@item -v
@itemx --verbose
@opindex -v
@opindex --verbose
Start in verbose mode, printing informational messages.
This is default for interactive mode.
@end table

@node Ftp commands
@section Commands interpreted by @command{ftp}

When @command{ftp} is awaiting commands from the user,
a prompt is displayed.
The default string is @samp{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 @command{ftp} itself.
Command names can be abbreviated to the shortest unique
string with identical beginning.

@table @code
@item ! [@var{command} [@var{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.

@item $ @var{macro-name} [@var{args}]
Execute the macro @var{macro-name} that was defined with the macdef
command.  Arguments are passed to the macro unglobbed.

@item account [@var{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.

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

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

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

@item 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.

@item bye
@itemx quit
Terminate the FTP session with the remote server and exit
@command{ftp}.  An end of file will also terminate the session and
exit.

@item case
Toggle the remote computer's use of letter case mapping during
@code{mget} commands.
When @code{case} is @samp{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 @samp{off},

@item cd @var{remote-directory}
Change the working directory on the remote machine to
@var{remote-directory}.

@item cdup
Change the remote machine's working directory to the parent of the
current working directory.

@item chmod @var{mode} @var{file-name}
Change the access permission of the file @var{file-name} on the remote
system to @var{mode}.

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

@item 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 @code{cr} is @samp{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 @code{cr} is @samp{off}.

@item delete @var{remote-file}
Delete the file @var{remote-file} on the remote machine.

@item debug [@var{debug-value}]
Toggle debugging mode.  If an optional @var{debug-value} is specified
it is used to set the debugging level.  When debugging is on,
@command{ftp} prints each command sent to the remote machine, preceded
by the string @samp{-->}.

@item dir [@var{remote-directory}] [@var{local-file}]
Print a listing of the contents in the directory
@var{remote-directory}, and, optionally, place the output in
@var{local-file}.  If interactive prompting is set, @command{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 @var{local-file} is a dash @samp{-},
then output is displayed on the terminal.

@item epsv4
Toggle the use of EPSV/EPRT for IPv4 addressing.  Default is off.

@item form @var{format}
Set the file transfer form to @var{format}.  The only supported
format is @samp{non-print}.

@item get @var{remote-file} [@var{local-file}]
@itemx recv @var{remote-file} [@var{local-file}]
Retrieve the @var{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 @code{case}, @code{ntrans}, and @code{nmap} settings.
The current settings for @code{type}, @code{form}, @code{mode},
and @code{structure} are effective during file transfer.

@item glob
Toggle file name expansion for @code{mdelete}, @code{mget}, and
@code{mput}.  If globbing is turned off with @code{glob}, the file
name arguments are taken literally and are not expanded.  Globbing for
@code{mput} is done as in @command{csh} syntax.
For @code{mdelete} and
@code{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 @samp{mls remote-files -}.

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

@item hash [@var{size}]
In the absence of an argument, toggle the state of hash-sign (@samp{#})
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.

@item help [@var{command}]
@itemx ? [@var{command}]
Print an informative message about the meaning of command.  If no
argument is given, @command{ftp} prints a list of the known commands.

@item idle [@var{seconds}]
Set the inactivity timer on the remote server to @var{seconds}
seconds.  If seconds is omitted, the current inactivity timer is
printed.

@item ipv4
Select IPv4 as the only addressing scheme.

@item ipv6
Select IPv6 as the only addressing scheme.

@item ipany
Allow IPv4 as well as IPv6 addressing.

@item lcd [@var{directory}]
Change the working directory on the local machine.  If no directory is
specified, the user's home directory is used.

@item lpwd
Print the name of the current working directory on the local machine.

@item ls [@var{remote-directory}] [@var{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 @command{ls -l} does.
Use @code{nlist} for a simple file listing.

If @var{remote-directory} is left unspecified, the current working
directory is used.  With interactive prompting set,
@command{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 @var{local-file} is a
dash@tie{}@samp{-}, then output is sent to the terminal.

@item macdef @var{macro-name}
Define a macro called @var{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
@var{macro-name} are significant when determining which
macro to execute.
Macros remain defined until a close command is executed.

The macro processor interprets @samp{$} and @samp{\} as
special characters.  A @samp{$} followed by a number (one or more
digits) is replaced by the corresponding argument on the macro's
invocation command line.
A @samp{$} followed by the letter @samp{i} tells the macro processor
that the macro is to perform a loop.
On the first pass, @samp{$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 @samp{\} followed by any character is replaced by that
character.  Use the backslash @samp{\} to prevent special treatment
of the dollar sign @samp{$}, as was just explained.

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

@item mdelete [@var{remote-files}]
Delete all @var{remote-files} on the remote machine.

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

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

@item mkdir @var{directory-name}
Make a directory on the remote machine.

@item mls @var{remote-files} @var{local-file}
Like @code{nlist}, except multiple remote files may be specified, and
the @var{local-file} must be specified.  If interactive prompting is
on, @command{ftp} will prompt the user to verify that the last
argument is the intended local file for storing output.
A dash @samp{-} is accepted as last argument without check!

@item mode [@var{mode-name}]
Set the file transfer mode to @var{mode-name}.  The default mode is
@samp{stream}, and it is also the only implemented mode.

@item modtime @var{file-name}
Show the last modification time of the file on the remote machine.

@item mput @var{local-files}
Consider the arguments to be local names and expand any wild card.
Execute a @code{put} for each file in the resulting list.
The remote file names are then computed by use of
@code{ntrans} and @code{nmap} settings.

@item newer @var{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 @code{get}.

@item nlist [@var{remote-directory}] [@var{local-file}]
Print a list of the files in a directory on the remote machine.  If
@var{remote-directory} is left unspecified, the current working
directory is used.  If interactive prompting is on, @command{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 @var{local-file} is @samp{-}, the output is sent to the
terminal.

@item nmap [@var{inpattern} @var{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 @code{mput} and
@code{put} commands issued without a specified remote target filename.
It as also applied to local file names during
@code{mget} and @code{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 @var{inpattern} and
@var{outpattern}.  The template @var{inpattern} is used on incoming
filenames (which may have already been processed according to the
@code{ntrans} and @code{case} settings).
Variable templating is accomplished
by including the sequences @samp{$1}, @samp{$2}, @dots{}, @samp{$9} in
@var{inpattern}.  Use @samp{\} to prevent this special treatment of
the character @samp{$}.  All other characters are treated literally,
and must be matched in a file name for @var{inpattern}
to bind substrings to variables.

For example, take a pattern @samp{$1.$2} and a file name
@file{mydata.data}.
Then @samp{$1} would have the value @samp{mydata}, and
@samp{$2} would be @samp{data}.

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

@example
nmap $1.$2.$3 [$1,$2].[$2,file]
@end example

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

Spaces may be included in @var{outpattern}, but are easily removed:

@example
nmap $1 |sed "s/ *$//" > $1
@end example

Use a backslash @samp{\} to escape the characters
@samp{$}, @samp{[}, @samp{]}, and @samp{,}.

@item ntrans [@var{inchars} [@var{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 @code{mput} commands and @code{put} commands
issued without a specified remote target filename.  If arguments are
specified, characters in local filenames are translated during
@code{mget} commands and @code{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 @var{inchars} are
replaced with the corresponding character in @var{outchars}.  If the
character's position in @var{inchars} is longer than the length of
@var{outchars}, the character is deleted from the file name.

@item open @var{host} [@var{port}]
@itemx open @var{user@@host} [@var{port}]
Establish a connection to the specified FTP server
at @var{host}.  An optional port number may be supplied,
in which case, @command{ftp} will attempt to contact the server
at that specific TCP port.  If the @code{autologin} option
is on (is so by default), @command{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 @var{user},
which otherwise is taken as identical to the user identity owning the
local session.

@item passive
Toggle passive mode.  If passive mode is turned on (default is off),
the @command{ftp} client will send a @code{PASV} command for all data
connections instead of the usual @code{PORT} command.  The @code{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 @code{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 @command{ftp} through a gateway
router or host that controls the directionality of traffic.  (Note
that though @command{ftp} servers are required to support the
@code{PASV} command by RFC 1123, some do not.)  If @command{epsv4}
has been set to on, the client will attempt @code{EPSV} before
@code{PASV} for IPv4.  As a last resort @code{LPSV} is attempted.
With IPv6 only @code{EPSV} and @code{LPSV} are possible.

@item 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
@code{mget} or @code{mput} will transfer all files, and any
@code{mdelete} will delete all files.

@item proxy @var{ftp-command}
Execute an @command{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 @code{open}, to establish the secondary control
connection.  Enter the command @code{proxy ?} to see other
commands usable for the secondary connection.  The following
commands behave differently when prefaced by @code{proxy}: @code{open}
will not define new macros during the auto-login process, @code{close}
will not erase existing macro definitions, @code{get} and @code{mget}
transfer files from the host on the primary control connection to the
host on the secondary control connection, and @code{put}, @code{mput},
and @code{append} transfer files from the host on the secondary
control connection to the host on the primary control connection.

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

@item put @var{local-file} [@var{remote-file}]
@itemx send @var{local-file} [@var{remote-file}]
Store a local file on the remote machine.  If @var{remote-file} is
left unspecified, the local file name is used after processing
according to any @code{ntrans} or @code{nmap} settings in naming the
remote file.  File transfer uses the current settings for type,
format, mode, and structure.

@item pwd
Print the name of the current working directory on the remote machine.

@item quote @var{arg}@dots{}
The arguments specified are sent, verbatim, to the remote FTP server.

@item reget @var{remote-file} [@var{local-file}]
@code{reget} acts like @code{get}, except that if @var{local-file}
exists and is smaller than @var{remote-file}, then @var{local-file} is
presumed to be a partially transferred copy of @var{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.

@item rhelp [@var{command-name}]
Request help from the remote FTP server.
If @var{command-name} is specified it is passed to the server as well.

@item rstatus [@var{file-name}]
With no arguments, show status of remote machine.  If filename is
specified, show status of @var{file-name} on remote machine.

@item rename [@var{from}] [@var{to}]
Rename the file @var{from} on the remote machine as @var{to}.
Name mapping takes effect without @var{to}.

@item 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.

@item restart @var{marker}
Restart the immediately following @code{get} or @code{put} at the
indicated marker.  On UNIX systems, @code{marker} is usually
a byte offset into the file.

@item rmdir @var{directory-name}
Delete a directory on the remote machine.

@item 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 @code{get} or @code{mget} command,
then a string @samp{.1} is appended to the name.
If the resulting name matches another existing file,
@samp{.2} is appended to the original name.  If this process continues
up to @samp{.99}, an error message is printed, and the transfer does
not take place.  The generated unique filename will be reported.  Note
that @code{runique} will not affect local files generated from a shell
command.  The default value is off.

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

@item site @var{arg}@dots{}
The arguments specified are sent, verbatim, to the remote FTP server
as a @code{SITE} command.

@item size @var{file-name}
Return size of @var{file-name} on remote machine.

@item status
Show the current status of @command{ftp}.

@item struct [@var{struct-name}]
Set the file transfer structure to @var{struct-name}.  By default
@samp{file} structure is used, which also is the only supported
value.

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

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

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

@item trace
Toggle packet tracing (feature is not implemented).

@item type [@var{type-name}]
Set the file transfer type to @var{type-name}.  If no type is
specified, the current type is printed.
The recognized type names are @samp{ascii}, @samp{binary},
@samp{ebcdic}, @samp{image}, and @samp{tenex}.
The default type is network ASCII.

@item umask [@var{newmask}]
Set the default umask on the remote server to @var{newmask}.  If
@var{newmask} is omitted, the current umask is printed.

@item user @var{user-name} [@var{password}] [@var{account}]
Identify yourself to the remote FTP server.  If the password is not
specified and the server requires it, @command{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
@command{ftp} is invoked with @code{auto-login} disabled, this process
is done automatically on initial connection to the FTP server.

@item 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.
@end table

Command arguments which have embedded spaces may be inclosed within
citation characters @samp{"}.

@node Ftp environment
@section Environment variables in use

@command{ftp} accesses the following environment variables.

@table @env
@item HOME
Used for locating a @file{.netrc} file, if one exists.

@item NETRC
Alternate location of the @file{.netrc} file,
taking precedence over the standard location.

@item SHELL
For determining the default shell interpreter.
@end table

@node Aborting a file transfer
@section Aborting a file transfer

To abort a file transfer, use the terminal interrupt key (usually
@kbd{C-c}).  Sending transfers will be immediately halted.
Receiving transfers will be halted by sending a FTP
protocol command @code{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 @code{ABOR} processing.  If the remote server does not support the
@code{ABOR} command, an @samp{ftp>} prompt will not appear until the
remote server has completed sending the requested file.

The terminal interrupt key sequence will be ignored when @command{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
@code{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
@command{ftp} program must be killed by hand.

@node File naming conventions
@section File naming conventions

Files specified as arguments to @command{ftp} commands are processed
according to the following rules.

@enumerate
@item
If the file name @samp{-} is specified, standard input (for reading)
or standard output (for writing) is used.

@item
If the first character of the file name is @samp{|}, the remainder of
the argument is interpreted as a shell command.  @command{ftp} then
forks a shell, using @code{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.  @samp{"ls -lt"}.

A particularly useful example of this mechanism in action, is

@example
ftp> dir . |less
@end example

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

@item
Failing the above checks, if @dfn{globbing} is enabled, local file
names are expanded according to the rules used by @command{csh};
c.f. the @code{glob} command.  If the @command{ftp} command expects a
single local file (e.g. @code{put}), only the first filename
generated by the globbing operation is used.

@item
For the commands @code{mget} and @code{get} with unspecified
local file name, the local file name is set to the remote file name,
which may be altered by a @code{case}, @code{ntrans}, or @code{nmap}
settings.
The resulting file name may then be modified if @code{runique} is set.

@item
For the commands @code{mput} and @code{put} with unspecified
remote file name, the remote file name is copied from the local
file name, which may be altered by a @code{ntrans} or @code{nmap}
settings.
The resulting file name may also be modified by the remote server if
@code{sunique} is set.
@end enumerate

@node File transfer parameters
@section File transfer parameters

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

@command{ftp} supports only the default values for the remaining file
transfer parameters: @code{mode}, @code{form}, and @code{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.

@node The .netrc file
@section The @file{.netrc} file
@flindex .netrc

The @file{.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 @env{NETRC}.
Both locations are overridden by the command line option @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:

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

@item 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:

@example
default login anonymous password user@@site
@end example

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

@item 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.

@item 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 @file{.netrc} file for any user other than anonymous,
@command{ftp} will abort the auto-login process if the @file{.netrc}
is readable by anyone besides the user.

@item 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 @code{ACCT} command if it does not.

@item macdef name
Define a macro.  This token functions like the @command{ftp}
@code{macdef} command functions.  A macro is defined with the
specified name; its contents begin with the next @file{.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.
@end table

@node rcp invocation
@chapter @command{rcp}: Copy files between machines
@pindex rcp

@command{rcp} copies files between machines.  Each file or directory
argument is either a remote file name of the form
@samp{rname@@rhost:path}, or a local file name (containing no @samp{:}
characters, or a @samp{/} before any @samp{:}s).

@noindent
Synopsis:

@example
rcp [@var{option}]@dots{} @var{old-file} @var{new-file}
rcp [@var{option}]@dots{} @var{files}@dots{} @var{directory}
@end example

@section Command line options
@anchor{rcp options}

@table @option
@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Use only IPv4.

@item -6
@itemx --ipv6
@opindex -6
@opindex --ipv6
Use only IPv6.

@item -d @var{directory}
@itemx --target-directory=@var{directory}
@opindex -d
@opindex --target-directory
Copy all source arguments into @var{directory}.

@item -f
@itemx --from
@opindex -f
@opindex --from
(Server mode only.) Copying from remote host.

@item -k @var{realm}
@itemx --realm=@var{realm}
@opindex -k
@opindex --realm
The option requests rcp to obtain tickets for the remote host in
realm @var{realm} instead of the remote host's realm.

@item -K
@itemx --kerberos
@opindex -K
@opindex --kerberos
Turns off all Kerberos authentication.

@item -p
@itemx --preserve
@opindex -p
@opindex --preserve
Causes @code{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 @code{umask} setting on the destination host.

@item -r
@itemx --recursive
@opindex -r
@opindex --recursive
If any of the source files are directories, @command{rcp} copies each
subtree rooted at that name; in this case the destination must be a
directory.

@item -t
@itemx --to
@opindex -t
@opindex --to
(Server mode only.) Copying to remote host.

@item -x
@itemx --encrypt
@opindex -x
@opindex --encrypt
Turns on encryption for all data passed via the @command{rcp} session.
This may impact response time and CPU utilization, but provides increased
security.

@end table

@command{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.

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

The destination user and hostname may have to be specified as
@samp{rhost.rname} when the destination machine is running the 4.2BSD
version of @command{rcp}.

@node rexec invocation
@chapter @command{rexec}: a remote execution program
@pindex rexec

@command{rexec} is a program that executes a program on another host.

@noindent
Synopsis:

@example
rexec --user=@var{login} --password=@var{pass} --host=@var{host}  \
@verb{|      |}[OPTION] @var{command}
@end example

@section Command line options
@anchor{rexec options}

@table @option
@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Use only IPv4 connections as all times.

@item -6
@itemx --ipv6
@opindex -6
@opindex --ipv6
Use only IPv6 connections.

@item -a
@itemx --ipany
@opindex -a
@opindex --ipany
Allow any address family for connections.  This is the default.

@item -e
@itemx --error=@var{port}
@opindex -e
@opindex --error
Specify the TCP port to use for stderr redirection, in case it is not
specified a random port will be used.

@item -h
@item --host=@var{name}
@opindex -h
@opindex --host
Specify the host with whom to connect: symbolic name or address.

@item -n
@itemx --noerr
@opindex -n
@opindex --noerr
If specified, an error stream will not be created.

@item -p
@itemx --password=@var{passwd}
@opindex -p
@opindex --password
Specify the password for logging-in.  The special value
consisting of a single dash @samp{-} will make @command{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.

@item -P
@itemx --port=@var{num}
@opindex -P
@opindex --port
Specify to which numerical port a connection shall be sought.
If it is not specified, then use port 512/tcp by default.

@item -u
@itemx --user=@var{name}
@opindex -u
@opindex --user
Specify the user with whom to log into the server.
@end table

@node rlogin invocation
@chapter @command{rlogin}: Remote login
@pindex rlogin

The @command{rlogin} command starts a terminal session on the
specified remote host, provided the required authentication
is successful.  The remote terminal type is the same as that
given in the @env{TERM} local environment variable.
The terminal and the window size stay the same, if the remote
host supports them, and any changes in size are transferred
as need may be.

When using the @command{rlogin} command, you can create a link
in your path, using a host name as the link name.  For example:

@example
# ln -s /usr/bin/rlogin @var{hostname}
# @var{hostname} -8
@end example

@noindent
Afterwards, the use of @var{hostname} will automatically invoke
@command{rlogin} to direct a log in request to the remote host
named @var{hostname}.

@command{rlogin} allows access to the remote host without the use of a
password.  The prerequisite is a suitable specification in @file{~/.rhosts}.
For details, @xref{rcmd, , rcmd, libc, The GNU C Library Reference Manual}.

@section Command line options
@anchor{rlogin options}

The options are as follows :

@table @option
@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Use only IPv4.

@item -6
@itemx --ipv6
@opindex -6
@opindex --ipv6
Use only IPv6.

@item -8
@itemx --8-bit
@opindex -8
@opindex --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 @kbd{C-S}/@kbd{C-Q}.

@item -d
@itemx --debug
@opindex -d
@opindex --debug
Turns on socket debugging on the TCP sockets used for communication
with the remote host.

@item -e @var{char}
@itemx --escape=@var{char}
@opindex -e
@opindex --escape
Allows user specification of the escape character, which is @samp{~}
by default.  This specification may be as a literal character, or as
an octal value in the form @samp{\nnn}.

@item -E
@itemx --no-escape
@opindex -E
@opindex --no-escape
Stops any character from being recognized as an escape character.
When used with the @option{-8} option, this provides a completely
transparent connection.

@item -l @var{user}
@itemx --user=@var{user}
@opindex -l
@opindex --user
By default, the remote username is the same as the local username.
This option, and the @samp{user@@host} format, allow the remote
user name to be made explicit, or changed.
@end table

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

@table @option
@item -k @var{realm}
@itemx --realm=@var{realm}
@opindex -k
@opindex --realm
The option requests rlogin to obtain tickets for the remote host in
realm @var{realm} instead of the remote host's realm.

@item -K
@itemx --kerberos
@opindex -K
@opindex --kerberos
Turns off all Kerberos authentication.

@item -x
@itemx --encrypt
@opindex -x
@opindex --encrypt
Turns on encryption for all data passed via the rlogin session.
This may impact response time and CPU utilization, but provides
increased security.
@end table

@section Escape characters and flow control

As long as the connection stands, the client program @command{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 @var{escape-char}.
The following sequences are supported:

@c An input line of the form of the two-character sequence
@itemize @bullet
@item
@kbd{@var{escape-char} .} disconnects from the remote host.

@item
@kbd{@var{escape-char} C-d} does the same.
(Termios character @samp{VEOF}.)

@item
@kbd{@var{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 @samp{VSUSP}.)

@item
@kbd{@var{escape-char} @var{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 @samp{VDSUSP}.)
@end itemize

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

All echoing takes place at the remote site, so that the @command{rlogin}
is transparent except possibly for transmission delays.
Flow control via @kbd{C-S} and @kbd{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 @code{iruserok} and @code{ruserok} functions
are used to authenticate the connection request, unless Kerberised
mode is in effect.  See the appropriate man pages for more information.

@section Kerberos Authentication

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

@node rsh invocation
@chapter @command{rsh}: Remote shell
@pindex rsh

@command{rsh} executes commands on a remote host and copies its local
standard input to that of the remote command, as well as the remote
standard output to the local standard output, and the remote standard
error to the local standard error.  Locally raised interrupt, quit and
terminate signals are all propagated to the remote command. Normally
@command{rsh} terminates when the remote command does so.

When using the @command{rsh} command, you can for convenience create
a link in your path, using a host name as name of the link.  For example:

@example
# ln -s /usr/bin/rsh @var{hostname}
# @var{hostname} ls
@end example

@noindent
Afterwards, @var{hostname} will be passed to @command{rsh} as host name
whenever the command @var{hostname} is issued.

@command{rsh} allows access to the remote host without the use of a
password.  The prerequisite is a suitable specification in @file{~/.rhosts}.
For details, @xref{rcmd, , rcmd, libc, The GNU C Library Reference Manual}.

If no command is specified for @command{rsh} ar argument following the
host name, then you will be logged in on the remote host using @command{rlogin}.

@section Command line options
@anchor{rsh options}

The options are as follows :

@table @option
@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Use only IPv4.

@item -6
@itemx --ipv6
@opindex -6
@opindex --ipv6
Use only IPv6.

@item -d
@itemx --debug
@opindex -d
@opindex --debug
Turns on socket debugging used for communication with the remote host.

@item -l @var{user}
@itemx --user=@var{user}
@opindex -l
@opindex --user
By default, the remote username is the same as the local username.
The @option{-l} option and the @samp{username@@host} format allow the
remote user name to be specified.  Kerberos authentication is used,
whenever available, and authorization is determined as in @command{rlogin}
(@pxref{rlogin invocation}).

@item -n
@itemx --no-input
@opindex -n
@opindex --no-input
Use @file{/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.
@end table

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

@table @option
@item -k @var{realm}
@itemx --realm=@var{realm}
@opindex -k
@opindex --realm
The option requests rsh to obtain tickets for the remote host in
realm @var{realm} instead of the remote host's realm.

@item -K
@itemx --kerberos
@opindex -K
@opindex --kerberos
Turns off all Kerberos authentication.

@item -x
@itemx --encrypt
@opindex -x
@opindex --encrypt
Turns on encryption for all data passed via the rsh session.  This
may impact response time and CPU utilization, but provides increased
security.
@end table

@noindent
Finally, some compatibility options are present:

@table @option
@item -8
@itemx --8-bit
@itemx -e @var{char}
@itemx --escape=@var{char}
@itemx -E
@itemx --no-escape
Ignored during normal operation, but passed on to @command{rlogin}
when @command{rsh} is invoked without a command argument.
@end table

@section 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:

@example
rsh otherhost  cat remotefile >> localfile
rsh otherhost  cat remotefile ">>" otherfile
@end example

@noindent
The first command appends the contents of @file{remotefile}, as found
on the remote host, to the file @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 @file{remotefile} to a file named @file{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

@example
cat remotefile >> localfile
@end example

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

@node talk invocation
@chapter @command{talk}: a communication program
@pindex talk

@command{talk} is a visual communication program which copies lines
from your terminal to that of another user.

@noindent
Synopsis:

@example
talk @var{person} [@var{ttyname}]
@end example

@section Invoking

The command line arguments are as follows:

@table @var
@item person
If you wish to talk to someone on your own machine, then @var{person}
is just the other person's login name.
If you wish to talk to a user on another host,
then @var{person} is of the form @samp{user@@host}.

@item ttyname
If you wish to talk to a local user who is logged in more than once,
the argument @var{ttyname} may be used to indicate the appropriate
terminal name, where @var{ttyname} typically is of the form @samp{ttyXX},
or @samp{pts/X}.
@end table

When first called, @command{talk} sends a message to
the addressed user:

@example
Message from TalkDaemon@@his_machine@dots{}
talk: connection requested by your_name@@your_machine.
talk: respond with: talk your_name@@your_machine
@end example

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

@example
talk @var{your_name}@@@var{your_machine}
@end example

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 @kbd{C-L} will cause the screen to be
reprinted, while erase, kill, and word kill characters will
behave normally.
In addition, @kbd{C-D} will cause both windows to be
locally cleared of all text.
This keystroke will appear as a simple @samp{^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 @kbd{C-C};
@command{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
@command{mesg} command.  It is system dependent whether
this message passing is enabled at the outset of a terminal session.
Certain commands, in particular @command{nroff} and @command{pr},
disable messages in order to prevent messy output.

@node telnet invocation
@chapter @command{telnet}: User interface to TELNET
@pindex telnet

Login to a remote system HOST, optionally using a (non-standard)
service port PORT.

@noindent
Synopsis:

@example
telnet [@var{OPTION}...] [@var{HOST} [@var{PORT}]]
@end example

@section Command line options
@anchor{telnet options}

@table @option
@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Use only IPv4.

@item -6
@itemx --ipv6
@opindex -6
@opindex --ipv6
Use only IPv6.

@item -8
@itemx --binary
@opindex -8
@opindex --binary
Use an 8-bit data path.

@item -a
@itemx --login
@opindex -a
@opindex --login
Attempt automatic login.

@item -b @var{address}
@itemx --bind=@var{address}
@opindex -b
@opindex --bind
Bind to specific local @var{address}.

@item -c
@itemx --no-rc
@opindex -c
@opindex --no-rc
Do not read the user's file @file{$HOME/.telnetrc}.

@item -d
@itemx --debug
@opindex -d
@opindex --debug
Turn on socket level debugging.

@item -e @var{char}
@itemx --escape=@var{char}
@opindex -e
@opindex --escape
Use @var{char} as escape character.

@item -E
@itemx --no-escape
@opindex -E
@opindex --no-escape
Do not use an escape character.

@item -k @var{realm}
@itemx --realm=@var{realm}
@opindex -k
@opindex --realm
Request Kerberos realm @var{realm} instead of whatever is
declared as default realm in the system's or user's settings.

@item -K
@itemx --no-login
@opindex -K
@opindex --no-login
Do not automatically login to the remote system.

@item -l @var{user}
@itemx --user=@var{user}
@opindex -l
@opindex --user
Attempt automatic login as @var{user}.

@item -L
@itemx --binary-output
@opindex -L
@opindex --binary-output
Use an 8-bit data path for output only.

@item -n @var{file}
@itemx --trace=@var{file}
@opindex -n
@opindex --trace
Record trace information into @var{file}.

@item -r
@itemx --rlogin
@opindex -r
@opindex --rlogin
Display a user-interface similar to that of @command{rlogin}.

@item -x
@itemx --encrypt
@opindex -x
@opindex --encrypt
If possible, encrypt the data stream.

@item -X @var{atype}
@itemx --disable-auth=@var{atype}
@opindex -X
@opindex --disable-auth
Disable authentication of type @var{atype}.
Use this option multiple times if more than one type
is to be disabled.  Standard choices are @samp{null},
@samp{kerberos_v4}, and @samp{kerberos_v5}.
@end table

@node tftp invocation
@chapter @command{tftp}: TFTP client
@pindex tftp

@command{tftp} is the user interface to the Internet TFTP, Trivial
File Transfer Protocol, which allows users to transfer files to and
from a remote machine.  The remote host may be specified on the
command line, in which case @command{tftp} uses host as the default
host for future transfers.

@noindent
Synopsis:

@example
tftp [@var{option}]@dots{} @var{host}
@end example

@section Commands

Once @command{tftp} is running, it issues the prompt and recognizes
the following commands:

@table @code
@item ? @var{command-name}
Print help information.

@item ascii
Shorthand for @code{mode ascii}

@item binary
Shorthand for @code{mode binary}

@item connect @var{host-name} [@var{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.

@item get @var{file-name}
@itemx get @var{remotename} @var{localname}
@itemx get @var{file}@dots{}
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 @samp{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.

@example
tftp> get [2001:1234::12]:issue
@end example

@item mode @var{transfer-mode}
Set the mode for transfers; @var{transfer-mode} may be one of
@samp{ascii} or @samp{binary}.  The default is @samp{ascii}.

@item put @var{file}
@itemx put @var{localfile} @var{remotefile}
@itemx put @var{file}@dots{} @var{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
@samp{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 @file{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 @command{get}.

@item quit
Exit @command{tftp}.  An end of file also exits.

@item rexmt @var{retransmission-timeout}
Set the per-packet retransmission timeout, in seconds.

@item status
Show current status.

@item timeout @var{total-transmission-timeout}
Set the total transmission timeout, in seconds.

@item trace
Toggle packet tracing.

@item verbose
Toggle verbose mode.
@end table

Because there is no user-login or validation within the @command{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.

@node inetd invocation
@chapter @command{inetd}: Internet super-server
@pindex inetd

@command{inetd} program should be run at boot time by /etc/rc.  It
then listens for connections on certain internet sockets.  When a
connection is found on one of its sockets, it decides what service the
socket corresponds to, and invokes a program to service the request.
The server program is invoked with the service socket as its standard
input, output and error descriptors.  After the program is finished,
inetd continues to listen on the socket (except in some cases which
will be described below).  Essentially, inetd allows running one
daemon to invoke several others, reducing load on the system.

There are two types of services that inetd can start: standard and
TCPMUX.  A standard service has a well-known port assigned to it; it
may be a service that implements an official Internet standard or is a
BSD-specific service.  As described in RFC 1078, TCPMUX services are
nonstandard services that do not have a well-known port assigned to
them.  They are invoked from inetd when a program connects to the
``tcpmux'' well-known port and specifies the service name.  This
feature is useful for adding locally-developed servers.

@menu
* Invocation::
* Configuration file::
* Built-in services::
* TCPMUX::
* Inetd Environment::
* Error Messages::
@end menu

@node Invocation
@section Invocation

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

@table @option
@opindex -d
@opindex --debug
@item -d
@itemx --debug
Turns on debugging.  With this option, @command{inetd} stays in
foreground and prints additional debugging information of stderr.

@item --environment
@opindex --environment
Pass local and remote socket information in environment variables.
@xref{Inetd Environment}.

@item --foreground
@opindex --foreground
Run in foreground mode.  Similar to debug mode, but does not enable
debug messages.

@item -p[@var{file}]
@itemx --pidfile[=@var{file}]
@opindex -p
@opindex --pidfile
Use @var{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.

@item --resolve
@opindex --resolve
Resolve IP addresses when setting environment variables.
@xref{Inetd Environment}.

@item -R @var{rate}
@itemx --rate=@var{rate}
@opindex --r
@opindex --rate
Specify the maximum number of times a service can be invoked in one
minute; the default is 1000.
@end table

@node Configuration file
@section Configuration file

Upon execution, inetd reads its configuration information from
configuration files and directories named on the command line.
By default these are @file{/etc/inetd.conf} and @file{/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):

@table @asis
@item [service node:]service name
The service-name entry is the name of a valid service in the file
@file{/etc/services}.  For ``internal'' services (@pxref{Built-in
services}), the service name must be the official name of the service
(that is, the first entry in @file{/etc/services}), or a numeric
representation thereof.  For TCPMUX services, the value of the
@samp{service name} field consists of the string @samp{tcpmux}
followed by a slash and the locally-chosen service name
(@pxref{TCPMUX}).

An optional @samp{service node} prefix is allowed for internet services.
When present, it supplies the local addresses @command{inetd} should
use when listening for that service.  @samp{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,
@command{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.:

@example
127.0.0.1,192.168.0.5:
@end example

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 @samp{*} 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.:

@example
*:
@end example

it causes any previous default address specifier to be forgotten.

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

@item protocol
The protocol must be a valid protocol as given in
@file{/etc/protocols}.  Examples might be @samp{tcp} or @samp{udp}.
TCPMUX services must use @samp{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 @samp{4} or @samp{6} to the protocol name.  For
example @samp{tcp4} will only accept IPv4 tcp connections and
@samp{udp6} will only accept IPv6 udp connections.

@item wait/nowait[.max]
The @samp{wait/nowait} entry specifies whether the server that is
invoked by @command{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 @samp{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''.  @command{comsat} and @command{talkd} are
both examples of the latter type of datagram server.  @command{tftpd} is an
example of a multi-threaded datagram server.

Servers using stream sockets generally are multi-threaded and use the
@samp{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 @samp{max} suffix (a decimal number),
e.g.: @samp{nowait.15}.

Stream-based servers that use @samp{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 @samp{nowait}.

@item 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.,
@samp{user:group} or @samp{user.group}.

@item 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 @samp{internal}.

It is common usage to specify @file{/usr/sbin/tcpd} in this field.

@item server program arguments
The server program arguments should be just as arguments normally are,
starting with @code{argv[0]}, which is the name of the program.  If
the service is provided internally, this entry must contain the word
@samp{internal}, or be empty.
@end table

@node Built-in services
@section Built-in services

The @command{inetd} program provides several ``trivial'' services
internally by use of routines within itself.  All these services can
operate both in @samp{stream} and in @samp{dgram} mode.  They are:

@table @asis
@item echo
Send back to the originating source any data received from it.  This
is a debugging and measurement tool.

@item discard
Silently throw away any data received.

@item chargen
This is a character generator service.  It can be operated as both
stream or dgram service.  When operating in @samp{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
@samp{dgram} mode, @command{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.

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

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

@node TCPMUX
@section TCPMUX
The TCPMUX protocol.

@quotation
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.
@end quotation

If the TCPMUX service name begins with a ``+'', @command{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 @samp{help} causes inetd to list TCPMUX
services in @file{inetd.conf}.

To define TCPMUX services, the configuration file must contain a
@samp{tcpmux internal} definition.

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

@example
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
@end example

@node Inetd Environment
@section Inetd Environment

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

@table @env
@item PROTO
Always @samp{TCP}.

@item TCPLOCALIP
Local IP address of the interface which accepted the connection.

@item TCPLOCALPORT
Port number on which the TCP connection was established.

@item TCPREMOTEIP
IP address of the remote client.

@item TCPREMOTEPORT
Port number on the client side of the TCP connection.
@end table

Additionally, if given the @option{--remote} option, @command{inetd}
sets the following environment variables:

@table @env
@item TCPLOCALHOST
DNS name of @env{TCPLOCALIP}.

@item TCPREMOTEHOST
DNS name of @env{TCPREMOTEIP}.
@end table

@node Error Messages
@section Error Messages

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

@table @samp
@item 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:

@enumerate 1
@item there are lots of hosts requesting the service within a short time period,
@item a ``broken'' client program is requesting the service too frequently,
@item a malicious user is running a program to invoke the service in a ``denial of service'' attack,
@item the invoked service program has an error that causes clients to retry quickly.
@end enumerate

Use the @option{-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.

@item service/protocol: No such user 'user', service ignored
@itemx 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.

@item service/protocol: No such user 'user', service ignored
@itemx 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.

@item service: can't set uid number
@itemx service: can't set gid number
The user or group ID for the entry's user is invalid.
@end table

@node syslogd invocation
@chapter @command{syslogd}: system service logging faclity
@pindex syslogd

@command{syslogd} is a system service that provides error logging
facility.  Messages are read from the UNIX domain socket
@file{/dev/log}, from an Internet domain socket specified in
@file{/etc/services}, and from the special device @file{/dev/klog} (to
read kernel messages).

@command{syslogd} creates the file @file{/var/run/syslog.pid}, and
stores its process id there.  This can be used to kill or reconfigure
@command{syslogd}.

The message sent to @command{syslogd} should consist of a single line.
The message can contain a priority code, which should be a preceding
decimal number in angle braces, for example, @code{<5>}.  This
priority code should map into the priorities defined in the include
file @code{sys/syslog.h}.

@example
syslogd [@var{options}]@dots{}
@end example

@table @option
@item -f @var{file}
@itemx --rcfile=@var{file}
@opindex -f
@opindex --rcfile
Override configuration (the default file is @file{/etc/syslog.conf}).

@item -D @var{dir}
@itemx --rcdir=@var{dir}
@opindex -D
@opindex --rcdir
Override configuration directory (the default is @file{/etc/syslog.d}).

@item -P @var{file}
@itemx --pidfile=@var{file}
@opindex -P
@opindex --pidfile
Override pidfile (the default file is @file{/var/run/syslogd.pid}).

@item -n
@itemx --no-detach
@opindex -n
@opindex --no-detach
Do not enter daemon mode.

@item -d
@itemx --debug
@opindex -d
@opindex --debug
Print debug information (implies @option{-n}).

@item -p @var{file}
@itemx --socket=@var{file}
@opindex -p
@opindex --socket
Override default UNIX domain socket @file{/dev/log}.

@item -a @var{socket}
Add UNIX socket to listen.  An unlimited number of sockets is allowed.

@item -r
@itemx --inet
@opindex -r
@opindex --inet
Receive remote messages via Internet domain socket.
Without this option no remote massages are received,
since there is no listening socket. Yet sockets for
forwarding are created on the fly as needed,
which might cause performance issues on busy systems.

@item -b @var{address}
@itemx --bind=@var{address}
@opindex -b
@opindex --bind
Restrict the listening Internet domain socket to a single address.
The default (given the use of @option{-r}) is a wildcard address,
implying that the server listens at every available address.
Any name will be resolved, and the lookup result will depend
on the options @option{-4}, @option{-6}, and @option{--ipany}.

@item --no-unixaf
@opindex --no-unixaf
Do not listen on UNIX domain sockets (overrides @option{-a} and
@option{-p}).

@item --no-klog
@opindex --no-klog
Do not listen to the kernel log device @file{/dev/klog}.

@item --ipany
@opindex --ipany
Allow both address families: IPv4 and IPv6.

@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Use only IPv4 for Internet domain sockets.

@item -6
@itemx --ipv6
@opindex -6
@opindex --ipv6
Use only IPv6 for Internet domain sockets.

@item --no-forward
@opindex --no-forward
Do not forward any messages (overrides @option{-h}).
This disables even temporary creation of forwarding
sockets, an ability which is otherwise active when
the option @option{-r} is left out.

@item -h
@itemx --hop
@opindex -h
@opindex --hop
Forward messages from remote hosts.

@item -m @var{interval}
@itemx --mark=@var{interval}
@opindex -m
@opindex --mark
Specify timestamp interval expressed in minutes (0 for no timestamping).

@item -l @var{hostlist}
@opindex -l
Log hosts in @var{hostlist} by their hostname.  Multiple lists are
allowed.

@item -s @var{domainlist}
@opindex -s
List of domains which should be stripped from the FQDN of hosts before
logging their name.  Multiple lists are allowed.

@item -T
@itemx --local-time
@opindex -T
@opindex --local-time
Ignore any time contained in a received message.
In its stead, record the time of reception on the local
system.  This circumvents problems caused by remote hosts
with skewed clocks.
@end table

@section Configuration file

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

Each configuration file consists of lines with two fields:
a @dfn{selector} field which specifies the
types of messages and priorities to which the line applies, and an
@dfn{action} field which specifies the action to be taken if a message
@command{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 @samp{\}.

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

A tagged selector commences with an exclamation mark,
as in @samp{!name}, or with a shebang, like @samp{#! name},
and continues with a program name, a @dfn{tag} in the sense
used by @command{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 @samp{*}.

The selector fields are encoded as a facility, followed by
a period (@samp{.}), 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: @samp{auth}, @samp{authpriv},
@samp{cron}, @samp{daemon}, @samp{ftp}, @samp{kern}, @samp{lpr}, @samp{mail},
@samp{mark}, @samp{news}, @samp{syslog}, @samp{user}, @samp{uucp} and
@samp{local0} through @samp{local7}.  These keywords (with the
exception of @samp{mark}) correspond to the similar @samp{LOG_} values
specified to the @samp{openlog} and @samp{syslog} library routines.
@xref{Syslog, , Syslog, libc, 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): @samp{emerg},
@samp{alert}, @samp{crit}, @samp{err, warning}, @samp{notice}, @samp{info}
and @samp{debug}.  These keywords correspond to the similar @samp{LOG_}
values specified to the syslog library routine.

@xref{syslog; vsyslog, , syslog and vsyslog, libc, 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 (@samp{;}) 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 (@samp{,}) characters.

An asterisk (@samp{*}) can be used to specify all facilities or all
levels.  Two asterisks (@samp{**}) 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 (@samp{=}) 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 (@samp{!}) 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 @samp{mark} receives a message at priority
@samp{info} every 20 minutes.  This is not enabled by a facility field
containing an asterisk.

The special level @samp{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:

@itemize @bullet
@item
A pathname (beginning with a leading slash).  Selected messages are
appended to the file.

You may prepend a minus (@samp{-}) 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.

@item
A named pipe, beginning with a vertical bar (@samp{|}) followed by a
pathname.  The pipe must be created with @command{mkfifo} before
@command{syslogd} reads its configuration file.  This feature is
especially useful for debugging.

@item
A hostname (preceded by an at (@samp{@@}) sign).  Selected messages
are forwarded to @command{syslogd} on the named host.

@item
A comma separated list of users.  Selected messages are written to
those users if they are logged in.

@item
An asterisk.  Selected messages are written to all logged-in users.

Blank lines and lines whose first non-blank character is a hash
(@samp{#}) character are ignored.
@end itemize

A configuration file might appear as follows:

@example
# 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
@end example

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

@node ftpd invocation
@chapter @command{ftpd}: FTP daemon
@pindex ftpd

@command{ftpd} is the Internet File Transfer Protocol server process.
The server uses the TCP protocol and listens at the port specified in
the @samp{ftp} service specification.

@example
ftpd [@var{option}]@dots{}
@end example

@table @option
@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Daemon uses only IPv4 addressing.  Ignored in inetd mode.

@item -6
@itemx --ipv6
@opindex -6
@opindex --ipv6
Daemon uses only IPv6 addressing.  Ignored in inetd mode.

@item -A
@itemx --anonymous-only
@opindex -A
@opindex --anonymous-only
Only anonymous login is allowed.

@item -a @var{auth}
@itemx --auth=@var{auth}
@opindex -a
@opindex --auth
Specify what authentication mechanism to use for incoming connections.
Possible values are: @samp{kerberos}, @samp{kerberos5}, @samp{opie},
@samp{pam}, and @samp{default}.

Anonymous logins will continue to work when this option is used,
unless the user @samp{ftp} is removed from the system.

@item  -D
@itemx --daemon
@opindex -D
@opindex --daemon
@command{ftpd} enters daemon-mode.  That allows @command{ftpd} to be
run without @command{inetd}.

@item -d
@itemx --debug
@opindex -d
@opindex --debug
Debugging information is written to the @code{syslog} using facility
@samp{LOG_FTP}.

@item -l
@itemx --logging
@opindex -l
@opindex --logging
Each successful and failed ftp session is logged using @code{syslog}
with a facility of @samp{LOG_FTP}.  If this option is specified twice,
the retrieve (@code{get}), store (@code{put}), append, delete, make
directory, remove directory and rename operations and their filename
arguments are also logged.

@item --non-rfc2577
@opindex --non-rfc2577
Do not follow the suggestion of RFC 2577 to suppress messages
that could help an attacker to conduct user name enumeration.
This option allows the server to return with an error message
immediately upon receipt of a user name.
Such information includes non-existence claims and expiration claims.
The ideal mode would otherwise be to fake the relevance of asking
for a password, and only thereafter report an invalid login.

@item -p @var{pidfile}
@itemx --pidfile=@var{pidfile}
@opindex -p
@opindex --pidfile
Change default location of @var{pidfile}.

@item -q
@itemx --no-version
@opindex -q
@opindex --no-version
Quiet mode.  No information about the version of the @command{ftpd} is
given to the client.

@item -T
@itemx --max-timeout
@opindex -T
@opindex --max-timeout
A client may also request a different timeout period; the maximum
period allowed may be set to timeout seconds with the @option{-T}
option.  The default limit is 2 hours.

@item -t @var{timeout}
@itemx --timeout=@var{timeout}
@opindex -t
@opindex --timeout
The inactivity timeout period is set to timeout seconds (the default
is 15 minutes).

@item -u @var{umask}
@itemx --umask=@var{umask}
@opindex -u
@opindex --umask
Set default umask, expressed in base 8.
@end table

The file @file{/etc/nologin} can be used to disable FTP access.  If
the file exists, @command{ftpd} displays it and exits.  If the file
@file{/etc/ftpwelcome} exists, @command{ftpd} prints it before issuing
the @samp{ready} message.  If the file @file{/etc/motd} exists,
@command{ftpd} prints it after a successful login.

If this server was compiled with PAM support, then any non-anonymous
connection request will also be checked for settings pertaining to
the PAM service @samp{ftp}, before finally being accepted.

Linux-PAM is particular in that it also provides a module
@samp{pam_ftp.so} influencing even anonymous access.
By convention the present server relies on the functionality
in that module when built on relevant systems.
However, the module is known to be partially broken since
ten years back, when one compares the claims in its manual page,
so not all claimed trickery is available!

@section Standards
@anchor{ftpd standards}

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

@multitable @columnfractions 0.3 0.7
@headitem Request  @tab  Description
@item ABOR         @tab  abort previous command
@item ACCT         @tab  specify account (ignored)
@item ALLO         @tab  allocate storage (vacuously)
@item APPE         @tab  append to a file
@item CDUP         @tab  change to parent of current working directory
@item CWD          @tab  change working directory
@item DELE         @tab  delete a file
@item EPSV         @tab  extended passive transfer request
@item EPRT         @tab  specify data connection port
@item HELP         @tab  give help information
@item LIST         @tab  give list files in a directory (``ls -lgA'')
@item LPRT         @tab  specify data connection port
@item LPSV         @tab  long passive transfer request
@item MKD          @tab  make a directory
@item MDTM         @tab  show last modification time of file
@item MODE         @tab  specify data transfer mode
@item NLST         @tab  give name list of files in directory
@item NOOP         @tab  do nothing
@item PASS         @tab  specify password
@item PASV         @tab  prepare for server-to-server transfer
@item PORT         @tab  specify data connection port
@item PWD          @tab  print the current working directory
@item QUIT         @tab  terminate session
@item REST         @tab  restart incomplete transfer
@item RETR         @tab  retrieve a file
@item RMD          @tab  remove a directory
@item RNFR         @tab  specify rename-from file name
@item RNTO         @tab  specify rename-to file name
@item SITE         @tab  non-standard commands
@item SIZE         @tab  return size of file
@item STAT         @tab  return status of server
@item STOR         @tab  store a file
@item STOU         @tab  store a file with a unique name
@item STRU         @tab  specify data transfer structure
@item SYST         @tab  show operating system type of server system
@item TYPE         @tab  specify data transfer type
@item USER         @tab  specify user name
@item XCUP         @tab  change to parent of current working directory (deprecated)
@item XCWD         @tab  change working directory (deprecated)
@item XMKD         @tab  make a directory (deprecated)
@item XPWD         @tab  print the current working directory (deprecated)
@item XRMD         @tab  remove a directory (deprecated)
@end multitable

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

@multitable @columnfractions 0.3 0.7
@item Request      @tab  Description
@item UMASK        @tab  change umask, e.g. @code{SITE UMASK 002}
@item IDLE         @tab  set idle-timer, e.g. @code{SITE IDLE 60}
@item CHMOD        @tab  change mode of a file, e.g. @code{SITE CHMOD0 0CHMOD1 1CHMOD2}
@item HELP         @tab  give help information.
@end multitable

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

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

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

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

@section Authentication
@anchor{ftpd authentication}

@command{ftpd} authenticates users according to four rules.

@enumerate
@item
The login name must be in the password data base, @file{/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.

@item
The login name must not appear in the file @file{/etc/ftpusers}.

@item
The user must have a standard shell.  If the file @file{/etc/shells}
exists and is readable, only programs listed there are considered standard
shells.  @command{ftpd} uses the C library function @code{getusershell}
to enumerate standard shells.  Standard shells are also known as valid
login shells, valid user shells, or permitted user shells.

@item
If the user name is @samp{anonymous} or @samp{ftp}, an anonymous ftp
account must be present in the password file (user @samp{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).
@end enumerate

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

In the case of anonymous access, @command{ftpd} takes special measures
to restrict the client's access privileges.  The server always performs
a chroot to the home directory of the @samp{ftp} user.

In order that system security is not breached,
it is recommended that the @samp{ftp} subtree be
constructed with care, following these rules:

@table @file
@item ~ftp
Make the home directory owned by @samp{root} and not writable by anyone.

@item ~ftp/bin
Make this directory owned by @samp{root} and not writable by anyone
(mode 555).  The program @command{ls} must be present to support the
list command, unless the server was compiled with libls support.
This program should be mode 111.

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

@item ~ftp/pub
Make this directory mode 777 and owned by @samp{ftp}.  Guests can then
place files which are to be accessible via the anonymous account in
this directory.
@end table

@section Configuration files
@anchor{ftpd files}

@table @samp
@item @file{/etc/ftpchroot}
List of users to enclose in a chrooted directory.
The anonymous user @samp{ftp} is always considered
to be a member of this list, explicit or not.

@item @file{/etc/ftpusers}
List of unwelcome/restricted users, always to be denied access.

@item @file{/etc/ftpwelcome}
Welcome notice printed before server identification
and any authentication exchange.

@item @file{/etc/motd}
Welcome notice presented after completed login.

@item @file{/etc/nologin}
If present, the contents are displayed and all further
access is refused.
@end table

@section File format of ftpusers and ftpchroot
@anchor{ftpusers file format}

The files @file{/etc/ftpusers} and @file{/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 @samp{#},
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 @samp{@@},
is treated as a wildcard and matches every input.

A field commencing with an at-sign @samp{@@} and then
continuing with an identifier, is understood as giving
the name of a group.
Should this name exist in @file{/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 @samp{@@} on a line by itself in
@file{/etc/ftpchroot}, will enforce chrooting
upon every user allowed to access the FTP service.
This gives a Draconian, protective configuration.

@node rexecd invocation
@chapter @command{rexecd}: server for @code{rexec}
@pindex rexecd

@command{rexecd} is the server for the @code{rexec} routine.  The
server provides remote execution facilities with authentication based
on user names and passwords.  It passes error messages and notices
to the @code{syslog} facility @samp{LOG_DAEMON}.

@example
rexecd [@var{option}]@dots{}
@end example

@command{rexecd} listens for service requests at the port indicated in
the @samp{exec} service specification.  When a service request is
received the following protocol is initiated:

@enumerate
@item
The server reads characters from the socket up to a NUL (@samp{\0})
byte.  The resultant string is interpreted as an ASCII number, base
10.

@item
If the number received in step 1 is non-zero, it is interpreted as the
port number of a secondary stream to be used for the stderr.  A second
connection is then created to the specified port on the client's
machine.

@item
A NUL terminated user name of at most 16 characters is retrieved on
the initial socket.

@item
A NUL terminated, unencrypted password of at most 16 characters is
retrieved on the initial socket.

@item
A NUL terminated command to be passed to a shell is retrieved on the
initial socket.  The length of the command is limited by the upper
bound on the size of the system's argument list.

@item
@command{rexecd} then validates the user as is done at login time and,
if the authentication was successful, changes to the user's home
directory, and establishes the user and group protections of the user.
If any of these steps fail the connection is aborted with a diagnostic
message returned.

@item
A NUL byte is returned on the initial socket and the command line is
passed to the normal login shell of the user.  The shell inherits the
network connections established by rexecd.
@end enumerate

@section Invoking

The only option is as follows:

@table @option
@item -l
@itemx --logging
@opindex -l
@opindex --logging
Raise logging level for this service; use more than once for
increased verbosity.  The @code{syslog} facility in use is
@samp{LOG_DAEMON}.

@end table

Should @command{rexecd} have been built with PAM support,
it reads any setting specified for a service named @samp{rexec}.

@section 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).

@table @samp
@item username too long
The name is longer than 16 characters.

@item password too long
The password is longer than 16 characters.

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

@item Login incorrect.
No password file entry for the user name existed.

@item Password incorrect.
The wrong password was supplied.

@item No remote directory.
The chdir command to the home directory failed.

@item Try again.
A fork by the server failed.

@item <shellname>: @dots{}
The user's login shell could not be started.  This message is returned
on the connection associated with the stderr, and is not ...
@c FIXME: Fill this out.
@end table

@emph{Note}, that indicating @samp{Login incorrect} as opposed to
@samp{Password incorrect} is a security breach which allows people to
probe a system for users with null passwords.

@node rlogind invocation
@chapter @command{rlogind}: Remote login server
@pindex rlogind

@command{rlogind} is the server for the @command{rlogin} client program
(@pxref{rlogin invocation}).  The server provides a remote login
facility with authentication based on privileged port numbers from
trusted hosts, or using authentication according to a Kerberos
protocol.

@command{rlogind} in daemon mode listens for service requests at the
port indicated in the @samp{login} service specification.  A common
alternative is to have the super-server @command{inetd} listen at
the same port, which then invokes @command{rlogind} as demand arises.
In Kerberised mode, the port is either @samp{eklogin}, or
@samp{klogin}, depending on preset encryption, or none.

The standard authentication procedure assumes the integrity of each
client machine and of the connecting medium.  This is insecure, since
it transmits credentials in clear text, but is useful in an ``open''
environment.  This weakness is reduced when running the service
in Kerberised version, at the price of a larger complexity of the
supporting infrastructure.  Using an encrypting Kerberised service
even avoids all clear text processing.

@section Invoking

The available options are as follows:

@table @option
@item -4
@itemx --ipv4
@opindex -4
@opindex --ipv4
Accept only IPv4 connections in daemon mode.

@item -6
@itemx --ipv6
@opindex -6
@opindex --ipv6
Only IPv6 connections in daemon mode.

@item -a
@itemx --verify-hostname
@opindex -a
@opindex --verify-hostname
Ask hostname for verification.

@item -d[@var{max}]
@itemx --daemon[=@var{max}]
@opindex -d
@opindex --daemon
Run in background daemon mode, optionally setting the maximal
number of simultaneously running client sessions.  The default
limit is 10.

@item -D[@var{level}]
@itemx --debug[=@var{level}]
@opindex -D
@opindex -debug
Set debug level, not implemented.

@item -l
@itemx --no-rhosts
@opindex -l
@opindex --no-rhosts
Ignore client's @file{.rhosts} file.

@item -L @var{name}
@itemx --local-domain=@var{name}
@opindex -L
@opindex --local-domain
Set local domain name, to which the server host belongs.  By default
the domain is recovered from the canonical name of the host.

@item -n
@itemx --no-keepalive
@opindex -n
@opindex --no-keepalive
Do not set SO_KEEPALIVE on sockets.  This decreases the ability
to close lost connections to once active clients.

@item -o
@itemx --allow-root
@opindex -o
@opindex --allow-root
Allow the root user to login, which is disallowed by default.

@item -p @var{port}
@itemx --port=@var{port}
@opindex -p
@opindex --port
Listen on given port.  Applicable only in daemon mode.

@item -r
@itemx --reverse-required
@opindex -r
@opindex --reverse-required
Require reverse resolvability of remote host's numerical IP.
@end table

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 @command{rlogind}
compiled with Kerberos support.

@table @option
@item -k
@itemx --kerberos
@opindex -k
@opindex --kerberos
Activate Kerberos authentication on all incoming requests.

@item -S @var{name}
@itemx --server-principal=@var{name}
@opindex -S
@opindex --server-principal
Set Kerberos server name, overriding canonical hostname.

@item -x
@itemx --encrypt
@opindex -x
@opindex --encrypt
Activate encryption of all data passed via the @command{rlogind} session.
This may impact response time and CPU utilization, but provides
increased security.  Only for Kerberised mode of operation.
@end table

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

@section Kerberos specific details

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

@command{rlogind} will, in Kerberised operation mode, as default
instantiate itself using the principal name
@samp{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 @option{--server-principal}, as follows:

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

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

@section Protocol details

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

@enumerate
@item
The server checks the client's source port.  If the port is not in the
range 512-1023, the server aborts the connection.

@item
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 @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.
@end enumerate

Once the source port and address have been checked, @command{rlogind}
proceeds
with the authentication process as described in @ref{rshd invocation}.
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 @command{login} program, invoked with the option @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 @samp{PTY} is invoked to
provide flow control using @kbd{C-S}/@kbd{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 @env{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 @option{-n} was in effect when starting @code{rlogind}.
The use of keepalive messages allows sessions to be timed out,
should the client crash, or otherwise become unreachable.

@xref{ruserok, , ruserok, libc, The GNU C Library Reference Manual},
for details.

@section Diagnostics

The exchange protocol states that a negotiation reaches a successful
completion as soon as the server @command{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:

@table @samp
@item Permission denied.
The client presented insufficient credentials,
or the client's address is not sufficiently resolvable
to pass the checks induced by options @option{-a} or @option{-r}.

@item Try again.
A fork by the server failed.
@end table

@node rshd invocation
@chapter @command{rshd}: Remote shell server
@pindex rshd

The @command{rshd} server is the server for the @code{rcmd} routine
and, consequently, for the @command{rsh} (@pxref{rsh invocation})
program.  The server provides remote execution facilities with
authentication based on privileged port numbers from trusted hosts.
The @command{rshd} server listens for service requests at the port
indicated in the @samp{cmd} service specification.  When a service
request is received the following protocol is initiated:

@enumerate
@item
The server checks the client's source port.  If the port is not in the
range 512--1023, the server aborts the connection.  However, this
condition is not applied for Kerberized service.

@item
The server reads characters from the socket up to a NUL (@samp{\0})
byte.  The resultant string is interpreted as an ASCII number, base
10.

@item
If the number received in step 2 is non-zero, it is interpreted as the
port number of a secondary stream to be used for the stderr.  A second
connection is then created to the specified port on the client's
machine.  The source port of this second connection is also in the
range 512--1023.

@item
The server checks the client's source address and requests the
corresponding host name.  If the hostname cannot be determined, the
dot-notation 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} option is
given, the addresses for the hostname are requested, verifying that
the name and address correspond.  If address verification fails, the
connection is aborted with the message, @samp{Host address mismatch.}

@item
A null terminated user name of at most 16 characters is retrieved on
the initial socket.  This user name is interpreted as the user
identity on the client's machine.

@item
A null terminated user name of at most 16 characters is retrieved on
the initial socket.  This user name is interpreted as a user identity
to use on the server's machine.

@item
A null terminated command to be passed to a shell is retrieved on the
initial socket.  The length of the command is limited by the upper
bound on the size of the system's argument list.

@item
Rshd then validates the user using @code{ruserok}, which uses the file
@file{/etc/hosts.equiv} and the @file{.rhosts} file found in the
user's home directory.  The @option{-l} option prevents @code{ruserok}
from doing any validation based on the user's @file{.rhosts} file,
unless the user is the superuser.

@item
If the file @file{/etc/nologin} exists and the user is not the
superuser, the connection is closed.

@item
A null byte is returned on the initial socket and the command line is
passed to the normal login shell of the user.  The shell inherits the
network connections established by @command{rshd}.

@item
Transport-level keepalive messages are enabled unless the @option{-n}
option is present.  The use of keepalive messages allows sessions to
be timed out if the client crashes or becomes unreachable.

@item
The @option{-L} option causes all successful accesses to be logged to
@command{syslogd} (@pxref{syslogd invocation}) as @samp{auth.info}
messages.
@end enumerate

@xref{ruserok, , ruserok, libc, The GNU C Library Reference Manual},
for details.

@section Invoking

The options are as follows:

@table @option
@item -a
@itemx --verify-hostname
@opindex -a
@opindex --verify-hostname
Ask hostname for verification.

@c @item -d
@c @itemx --daemon
@c @opindex -d
@c @opindex --daemon
@c Daemon mode.

@item -k
@itemx --kerberos
@opindex -k
@opindex --kerberos
Use Kerberos authentication.

@item -l
@itemx --no-rhosts
@opindex -l
@opindex --no-rhosts
Ignore @file{.rhosts} file.

@item -L
@itemx --log-sessions
@opindex -L
@opindex --log-sessions
Log successful logins.

@item -n
@itemx --no-keepalive
@opindex -n
@opindex --no-keepalive
Do not set SO_KEEPALIVE.

@item -S @var{name}
@itemx --servername=@var{name}
@opindex -S
@opindex --servername
Set Kerberos server name, overriding canonical hostname.

@item -v
@itemx --vacuous
@opindex -v
@opindex --vacuous
Fail any call asking for non-Kerberos authentication.

@c OBSOLETE?
@c @item -x
@c @itemx --encrypt
@c @opindex -x
@c @opindex --encrypt
@c Turns on DES encryption for all data passed via the @command{rshd}
@c session.  This may impact response time and CPU utilization, but
@c provides increased security.

@c @item -D[@var{level}]
@c @itemx --debug[=@var{level}]
@c @opindex -D
@c @opindex -debug
@c Set debug level, not implemented.

@c @item -o
@c @itemx --allow-root
@c @opindex -o
@c @opindex --allow-root
@c Allow uid == 0 to login, disabled by default

@c @item -p @var{port}
@c @itemx --port=@var{port}
@c @opindex -p
@c @opindex --port
@c Listen on given port (valid only in daemon mode).

@item -r
@itemx --reverse-required
@opindex -r
@opindex --reverse-required
Demand that the client's IP address be resolvable
as a host name.
@end table

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

@section 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).

@table @samp
@item Locuser too long
The name of the user on the client's machine is longer than 16
characters.

@item Ruser too long
The name of the user on the remote machine is longer than 16
characters.

@item Command too long
The command line passed exceeds the size of the argument list (as
configured into the system).

@item Login incorrect
No password file entry for the user name existed.

@item Remote directory
The chdir command to the home directory failed.

@item Permission denied
The authentication procedure described above failed,
or address resolution was insufficient.

@item Can't make pipe.
The pipe needed for the stderr, wasn't created.

@item Can't fork; try again.
A fork by the server failed.

@item <shellname>: @dots{}
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.
@end table

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.

@node talkd invocation
@chapter @command{talkd}: a server for communication between users
@pindex talkd

@command{talkd} is a server that notifies users that someone else
wants to initiate a conversation.  It acts as a repository of
invitations, responding to requests by clients wishing to rendezvous
for a conversation.

This implementation uses the newer protocol @samp{ntalk/udp},
and is intended to be invoked by a super-server
@command{inetd} at that datagram port.
It is recommended that @command{inetd} launch @command{talkd}
with ownership @samp{nobody:tty}, or with @samp{tty:tty}.
However, this works with ACL only if @file{.talkrc} can be assumed
to be world readable for all users.
This failing, the process ownership will need to be @samp{root:tty}
if the ACL-mechanism is to be usable and trustworthy.

Keep in mind that this service is usable with IPv4 only,
since the exchange protocol was conceived to handle only
this particular address family.
This fact is independent of the abilities of @command{inetd}.

Observe also that the server @command{talkd} depends
on the name returned by @command{hostname}, for establishing
connections between interested parties.
A server @command{talkd} running on a multi-homed host
is not able to respond to invitations for a valid host name
that differs from the name reported by @command{hostname}.

The present implementation offers ACL-mechanisms for fine
grained access control.

@section Invoking

The following switches and options are available.

@table @option
@item -a @var{file}
@itemx --acl=@var{file}
@opindex -a
@opindex --acl
Read site-wide ACLs from @var{file}.

@item -d
@itemx --debug
@opindex --debug
@opindex --d
Enable debugging.

@item -i @var{seconds}
@itemx --idle-timeout=@var{seconds}
@opindex -i
@opindex --idle-timeout
Set idle timeout length

@item -l
@itemx --logging
@opindex -i
@opindex --logging
Enable a somewhat enhanced logging verbosity, reporting
attempted and dropped connections, as well as
some more unexpected events that might arise.

@item -r @var{seconds}
@itemx --request-ttl=@var{seconds}
@opindex -r
@opindex --request-ttl
Set time-to-live length for requests.

@item -S
@itemx --strict-policy
@opindex -S
@opindex --strict-policy
Apply strict ACL policy on this system. This means that
the site-wide ACL must provide explicit @samp{allow}
rules for admitting traffic at all.

@item -t @var{seconds}
@itemx --timeout=@var{seconds}
@opindex -t
@opindex --timeout
Set timeout length.
@end table

@section Modus operandi

In normal operation, a client, the caller, initiates a rendezvous by
sending a @code{CTL_MSG} of type @samp{LOOK_UP} to the server (see
@file{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 @samp{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 @file{.talkrc} is parsed, if at all present, in order to even
further fine tune access to this particular user.

@section Access control in talkd

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

The format of this access control list is shared with the user specific
file @file{.talkrc}.  Normally the site-wide setting operates with
a default value @samp{allow}, but specifying the option @option{-S},
or @option{--strict-policy}, changes this default action to @samp{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

@example
action user-exp [net-exp @dots{}]
@end example

@noindent
Each active line must contain at least two fields:
an @code{action} and a @code{user-exp}.

The first field, @code{action}, must be either of @samp{allow} and @samp{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, @code{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 @option{-a}, the expression @code{user-exp} is matched
against the requested local user name, that of the callee.

While checking the callee's private ACL-file @file{.talkrc},
the matching of @code{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 @code{net-exp}. Each of these is either the simple
word @samp{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 @samp{any}.
The absence of a net list is equivalent to specifying
a single @samp{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 @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 @code{net-exp} matches the incoming IPv4 address,
then the regular expression @code{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 @samp{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 @samp{deny}, from a system wide ACL,
will be used as the final action.

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

@node telnetd invocation
@chapter @command{telnetd}: Telnet server
@pindex telnetd

@example
telnetd [@var{option}]@dots{}
@end example

@table @option
@item --accept-env @var{VAR}
@opindex --accept-env
Allow clients to define the @var{VAR} environment variable.  GNU
@command{telnetd} removes all environment variables by default since
many of them can be used to escalate privileges.

@item -a @var{authmode}
@itemx --authmode=@var{authmode}
@opindex -a
@opindex --authmode
Specify what mode to use for authentication.  Allowed values are:
@samp{none}, @samp{other}, @samp{user}, @samp{valid}, and @samp{off}.

@item -E @var{string}
@itemx --exec-login=@var{string}
@opindex -E
@opindex --exec-login
Set program to be executed instead of @command{/bin/login}.

@item -h
@itemx --no-hostinfo
@opindex -h
@opindex --no-hostinfo
Do not print host information before login has been completed.

@item -l[@var{mode}]
@itemx --linemode=[@var{mode}]
@opindex -l
@opindex --linemode
Set line mode.  An empty argument will force line read mode at all times.
The only recognised value is otherwise @samp{nokludge}.

@item -n
@itemx --no-keepalive
@opindex -n
@opindex --no-keepalive
Disable TCP keep-alive.

@item -S @var{principal}
@itemx --server-principal=@var{principal}
@opindex -S
@opindex --server-principal
Set principal name for the server, to be used in Kerberos
authentication.  The value @var{principal} can be set
to provide full specification like @samp{srv.local@@REALM}
and @samp{tnt/localhost@@REALM}, where the first uses the
standard prefix `host/'. Or @var{principal} can override
default settings in part only, like @samp{srv.local},
@samp{tnt/srv.local}, or @samp{@@REALM}.

@item -U
@itemx --reverse-lookup
@opindex -U
@opindex --reverse-lookup
Refuse connections from addresses that cannot be mapped back into a
symbolic name.
A client is accepted only if the IP address can be resolved as
a host name, and the same name is resolvable to addresses among
which the clients's address is included.

@item -X @var{authtype}
@itemx --disable-auth-type=@var{authtype}
@opindex -X
@opindex --disable-auth-type
Disable the use of the given authentication type.
Use this option multiple times if more than one type
is to be disabled.  Standard choices are @samp{null},
@samp{kerberos_v4}, and @samp{kerberos_v5}.
@end table

@section Crafting an execution string

The server @command{telnetd} contains a built-in execution string
which invokes @command{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 @option{-E},
or @option{--exec-login}, to override the built-in execution of
@command{login}, thus allowing almost any choice of handler.

@noindent
A custom execution string could look like

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

@noindent
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, @command{telnetd} offers some replacement
tokens that dynamically are replaced by content.  All are of
the form @code{%<var>}, where @samp{<var>} is a single letter
from the following collection of selectors.
A valid letter is called @dfn{variable}.
The mark @emph{conditional}, appearing below, indicates that the
corresponding variable is conditionally assigned a value.

@table @code
@item %a
@opindex %a
Returns @samp{ok} whenever authentication is complete. @emph{conditional}

@item %d
@opindex %d
Produces a time and date string.

@item %h
@opindex %h
Gives the remote host name in canonical form.

@item %l
@opindex %l
States the local host name, also in canonical form.

@item %L
@opindex %L
Returns the path of the pseudo terminal assigned to the client.

@item %t
@opindex %t
Gives the terminal device stripped of the leading @samp{/dev/}.

@item %T
@opindex %T
States the terminal type, like @samp{xterm}. @emph{conditional}

@item %u
@opindex %u
Provides the authenticated user name. @emph{conditional}

@item %U
@opindex %U
Returns the user name passed as an environment variable @env{USER}
by the remote client software.  The value is empty, should the
environment not provide a value.

@end table

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

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

@noindent
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

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

@noindent
In case authentication was completed as user @samp{sigge},
the execution string would resolve to

@example
/usr/local/sbin/avrop  pts/1 sigge krb5
@end example

@noindent
In all other cases the result would be

@example
/usr/local/sbin/avrop  pts/1 $USER
@end example

@noindent
where @code{$USER} is the value of the corresponding environment
variable and could possibly be empty.

@node tftpd invocation
@chapter @command{tftpd}: TFTP server
@pindex tftpd

@command{tftpd} is intended to be invoked via @command{inetd}
at all times.

@noindent
Synopsis:

@example
tftpd [@var{options}] [@var{directory} @dots{}]
@end example

@table @option
@item -g @var{group}
@itemx --group=@var{group}
@opindex -g
@opindex --group
Specify group membership of the process owner.
This is used only along with the option @option{-s},
and replaces the group membership that comes from
the process owner himself.

@item -l
@itemx --logging
@opindex -l
@opindex --logging
Enable logging.

@item -n
@itemx --nonexistent
@opindex -n
@opindex --nonexistent
Suppress negative acknowledgement of requests for nonexistent relative
filenames.

@item -s @var{dir}
@itemx --secure-dir=@var{dir}
@opindex -s
@opindex --secure-dir
Let the serving process change its root directory to @var{dir}
before attending to any requests.
This directory is not observable by any client, but improves
server isolation, since servable contents must be located
below this chrooted directory @var{dir}.

@item -u @var{user}
@itemx --user=@var{user}
@opindex -u
@opindex --user
Specify the process owner for serving requests.
Only relevant along with the option @option{-s}.
The default name is @samp{nobody}.
@end table

@section Directory prefixes
@anchor{tftpd validation}

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

@itemize @bullet
@item
Relative pathnames are ignored.

@item
At most twenty prefixes are approved, the rest is discarded.
@end itemize

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

@itemize @bullet
@item
Every file request containing the substring @samp{/../} is denied,
as is a file name beginning with @samp{../}.

@item
Write requests must specify absolute locations.

@item
A file request, if specified as an @emph{absolute} pathname,
must begin with one of the approved directory prefixes,
should at least one such prefix have been accepted.

@item
In the absence of a prefix collection, any absolute pathname is
accepted, should the corresponding file exist.

@item
A file request, if specified as a @emph{relative} name,
will only be searched for below the acceptable prefixes,
should at least one such prefix have been approved.

@item
A request for a relatively named file, is denied in the absence
of approved directory prefixes.

@item
The resulting file must be world readable, or world writable,
for a read request, or a write request, to succeed.
@end itemize

@section Use cases
@anchor{tftpd setup cases}

The standard use case is an entry in @file{/etc/inetd.conf} like

@example
tftp dgram udp4 wait root /usr/sbin/tftpd \
@verb{        } tftpd /tftpboot /altboot
@end example

@noindent
This would allow the TFTP client to use any of

@example
get kernel
get /tftpboot/kernel
get kernel.alt
get /altboot/kernel.alt
get /etc/motd
@end example

@noindent
given that @file{/tftpboot/kernel} and @file{/altboot/kernel.alt} exist.
Observe that also @file{/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

@example
tftp dgram udp4 wait root /usr/sbin/tftpd \
@verb{        } tftpd --secure-dir=/srv/tftp-root  /tftpboot /altboot
@end example

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

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

@node uucpd invocation
@chapter @command{uucpd}: Unix to Unix Copy relay daemon.
@pindex uucpd

@command{uucpd} is a relay daemon responsible for accepting
TCP transported connections for @command{uucico}.  It is started
by @command{inetd}, conducts any authentication, and then hands
acceptable requests over to @command{uucico}.

@example
uucpd [@var{option}]...
@end example

@section Options

There is a single, specific option available:

@table @option
@item -u @var{location}
@itemx --uucico=@var{location}
@opindex -u
@opindex --uucico
Replace the hard coded location of @command{uucico} with the
value specified as @var{location}.
@end table

@section 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 @command{uucpd}
verifies that the stated user also has user shell location
identical to the full file system location of @command{uucico}.
Should that not be the case, the request is declined.

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

@node GNU Free Documentation License
@appendix GNU Free Documentation License
@include fdl-1.3.texi

@page
@node Index
@unnumbered Index

@printindex cp

@bye