HardenedBSD/usr.sbin/tcpsso/tcpsso.8
Michael Tuexen 2edac5a60a tcpsso: improve man page
Address comment provided by bcr.

MFC after:		1 week
Sponsored by:		Netflix, Inc.
2024-08-03 20:30:10 +02:00

260 lines
5.6 KiB
Groff

.\"
.\" SPDX-License-Identifier: BSD-2-Clause
.\"
.\" Copyright (c) 2022 Michael Tuexen <tuexen@FreeBSD.org>
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.Dd August 3, 2024
.Dt TCPSSO 8
.Os
.Sh NAME
.Nm tcpsso
.Nd set a socket option on a TCP endpoint
.Sh SYNOPSIS
.Nm
.Fl i Ar id
.Op Ar level
.Ar optname
.Ar optval
.Nm
.Fl a
.Op Ar level
.Ar optname
.Ar optval
.Nm
.Fl C Ar cc-algo
.Op Fl S Ar stack
.Op Fl s Ar state
.Op Ar level
.Ar optname
.Ar optval
.Nm
.Op Fl C Ar cc-algo
.Fl S Ar stack
.Op Fl s Ar state
.Op Ar level
.Ar optname
.Ar optval
.Nm
.Op Fl C Ar cc-algo
.Op Fl S Ar stack
.Fl s Ar state
.Op Ar level
.Ar optname
.Ar optval
.Sh DESCRIPTION
The
.Nm
command applies a
.Ar level
level socket option with name
.Ar optname
and value
.Ar optval
on a TCP endpoint from the command line.
.Pp
TCP endpoints in the
.Dv SYN_RCVD
state can only be handled by
.Nm
if their prior state was
.Dv SYN_SENT .
.Pp
.Op Ar level
can be specified as a non-negative number or a symbolic name like
.Dv SOL_SOCKET ,
.Dv IPPROTO_IP ,
.Dv IPPROTO_IPV6 ,
or
.Dv IPPROTO_TCP .
If not specified,
.Nm
deduces the level from
.Ar optname ,
if provided as a symbolic name.
If that is not the case,
.Dv IPPROTO_TCP
is used.
.Pp
.Ar optname
can be specified as a non-negative number or a symbolic name like
.Dv SO_DEBUG ,
.Dv IP_TOS ,
.Dv IPV6_TCLASS ,
.Dv TCP_LOG ,
.Dv TCP_CONGESTION ,
or
.Dv TCP_FUNCTION_BLK .
.Pp
.Ar optval
can be an integer value, which will be converted to a binary value and
passed as an int value.
If it cannot be parsed as an integer value, it will be processed as a string.
If the
.Ar optname
is
.Dv TCP_FUNCTION_BLK
then
.Ar optval
is converted to a
.Vt "struct tcp_function_set" .
.Pp
If
.Fl i Ar id
is specified then
.Nm
will apply the socket option to the TCP endpoint with the
.Dv inp_gencnt
provided as
.Ar id .
The
.Dv inp_gencnt
for existing TCP endpoints can be determined by using
.Xr sockstat 1 .
.Pp
If
.Fl a
is specified then
.Nm
will apply the socket option to all TCP endpoints subject to the above state
restrictions.
.Pp
If
.Fl C Ar cc-algo
is specified then
.Nm
will apply the socket option to all TCP endpoints using the TCP
congestion control algorithm
.Ar cc-algo
and subject to the above state restrictions.
.Pp
If
.Fl S Ar stack
is specified then
.Nm
will apply the socket option to all TCP endpoints using the TCP
stack
.Ar stack
and subject to the above state restrictions.
.Pp
If
.Fl s Ar state
is specified then
.Nm
will apply the socket option to all TCP endpoints being in the state
.Ar state .
.Ar state
is one of
.Dv CLOSED ,
.Dv LISTEN ,
.Dv SYN_SENT ,
.Dv SYN_RCVD ,
.Dv ESTABLISHED ,
.Dv CLOSE_WAIT ,
.Dv FIN_WAIT_1 ,
.Dv CLOSING ,
.Dv LAST_ACK ,
.Dv FIN_WAIT_2,
or
.Dv TIME_WAIT .
Using
.Dv SYN_RCVD
only applies to TCP endpoints in the state
.Dv SYN_RCVD
if their prior state was
.Dv SYN_SENT .
.Pp
If multiple of
.Fl C Ar cc-algo ,
.Fl S Ar stack ,
and
.Fl s Ar state
are specified,
.Nm
will apply the socket option to all TCP endpoints using the congestion
control algorithm
.Ar cc-algo ,
being in the state
.Ar state ,
and using the TCP stack
.Ar stack ,
if specified.
.Pp
If none of the
.Fl a ,
.Fl C ,
.Fl S ,
or
.Fl s
options are specified then the option
.Fl i
must be specified.
.Sh EXIT STATUS
.Ex -std
.Sh EXAMPLES
To diagnose a problem with a particular TCP connection to
.Xr sshd 8 ,
first determine its
.Dv inp_gencnt
using
.Xr sockstat 1 :
.Bd -literal -offset indent
# sockstat -4 -c -i -p 22 -P tcp -q
root sshd 827 4 tcp4 \e
192.168.1.1:22 192.168.1.2:53736 435
.Ed
.Pp
Then, use the following command to enable Black Box Logging on it:
.Bd -literal -offset indent
# tcpsso -i 435 TCP_LOG 4
.Ed
.Pp
To switch all TCP endpoints from using the freebsd stack to the RACK stack use:
.Bd -literal -offset indent
# tcpsso -S freebsd TCP_FUNCTION_BLK rack
.Ed
.Pp
Switch all TCP endpoints in the LISTEN state to the RACK stack by using:
.Bd -literal -offset indent
# tcpsso -s LISTEN TCP_FUNCTION_BLK rack
.Ed
.Pp
The following command will set the congestion control module of all TCP
endpoints currently using cubic as its congestion control algorithm to the
congestion control algorithm newreno:
.Bd -literal -offset indent
# tcpsso -C cubic TCP_CONGESTION newreno
.Ed
.Sh SEE ALSO
.Xr sockstat 1 ,
.Xr setsockopt 2 ,
.Xr tcp 4 ,
.Xr tcp_functions 9
.Sh HISTORY
The
.Nm
command first appeared in
.Fx 14 .
.Sh AUTHORS
.An Michael Tuexen Aq Mt tuexen@FreeBSD.org