xref: /freebsd/share/man/man4/mod_cc.4 (revision af23369a6deaaeb612ab266eb88b8bb8d560c322)
1.\"
2.\" Copyright (c) 2010-2011 The FreeBSD Foundation
3.\"
4.\" This documentation was written at the Centre for Advanced Internet
5.\" Architectures, Swinburne University of Technology, Melbourne, Australia by
6.\" David Hayes and Lawrence Stewart under sponsorship from the FreeBSD
7.\" Foundation.
8.\"
9.\" Redistribution and use in source and binary forms, with or without
10.\" modification, are permitted provided that the following conditions
11.\" are met:
12.\" 1. Redistributions of source code must retain the above copyright
13.\"    notice, this list of conditions and the following disclaimer.
14.\" 2. Redistributions in binary form must reproduce the above copyright
15.\"    notice, this list of conditions and the following disclaimer in the
16.\"    documentation and/or other materials provided with the distribution.
17.\"
18.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
19.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
22.\" ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28.\" SUCH DAMAGE.
29.\"
30.\" $FreeBSD$
31.\"
32.Dd September 13, 2022
33.Dt MOD_CC 4
34.Os
35.Sh NAME
36.Nm mod_cc
37.Nd Modular congestion control
38.Sh DESCRIPTION
39The modular congestion control framework allows the TCP implementation to
40dynamically change the congestion control algorithm used by new and existing
41connections.
42Algorithms are identified by a unique
43.Xr ascii 7
44name.
45Algorithm modules can be compiled into the kernel or loaded as kernel modules
46using the
47.Xr kld 4
48facility.
49.Pp
50The default algorithm is CUBIC, and all connections use the default unless
51explicitly overridden using the
52.Dv TCP_CONGESTION
53socket option (see
54.Xr tcp 4
55for details).
56The default can be changed using a
57.Xr sysctl 3
58MIB variable detailed in the
59.Sx MIB Variables
60section below.
61.Pp
62Algorithm specific parameters can be set or queried using the
63.Dv TCP_CCALGOOPT
64socket option (see
65.Xr tcp 4
66for details).
67Callers must pass a pointer to an algorithm specific data, and specify
68its size.
69.Pp
70Unloading a congestion control module will fail if it is used as a
71default by any Vnet.
72When unloading a module, the Vnet default is
73used to switch a connection to an alternate congestion control.
74Note that the new congestion control module may fail to initialize its
75internal memory, if so it will fail the module unload.
76If this occurs often times retrying the unload will succeed since the temporary
77memory shortage as the new CC module malloc's memory, that prevented the
78switch is often transient.
79.Sh MIB Variables
80The framework exposes the following variables in the
81.Va net.inet.tcp.cc
82branch of the
83.Xr sysctl 3
84MIB:
85.Bl -tag -width ".Va hystartplusplus.css_growth_div"
86.It Va available
87Read-only list of currently available congestion control algorithms by name.
88.It Va algorithm
89Returns the current default congestion control algorithm when read, and changes
90the default when set.
91When attempting to change the default algorithm, this variable should be set to
92one of the names listed by the
93.Va net.inet.tcp.cc.available
94MIB variable.
95.It Va abe
96Enable support for RFC 8511,
97which alters the window decrease factor applied to the congestion window in
98response to an ECN congestion signal.
99Refer to individual congestion control man pages to determine if they implement
100support for ABE and for configuration details.
101.It Va abe_frlossreduce
102If non-zero, apply standard beta instead of ABE-beta during ECN-signalled
103congestion recovery episodes if loss also needs to be repaired.
104.It Va hystartplusplus.bblogs
105This boolean controls if black box logging will be done for hystart++ events.
106If set to zero (the default) no logging is performed.
107If set to one then black box logs will be generated on all hystart++ events.
108.It Va hystartplusplus.css_rounds
109This value controls the number of rounds that CSS runs for.
110The default value matches the current internet-draft of 5.
111.It Va hystartplusplus.css_growth_div
112This value controls the divisor applied to slowstart during CSS.
113The default value matches the current internet-draft of 4.
114.It Va hystartplusplus.n_rttsamples
115This value controls how many rtt samples must be collected in each round for
116hystart++ to be active.
117The default value matches the current internet-draft of 8.
118.It Va hystartplusplus.maxrtt_thresh
119This value controls the maximum rtt variance clamp when considering if CSS is needed.
120The default value matches the current internet-draft of 16000 (in microseconds).
121For further explanation please see the internet-draft.
122.It Va hystartplusplus.minrtt_thresh
123This value controls the minimum rtt variance clamp when considering if CSS is needed.
124The default value matches the current internet-draft of 4000 (in microseconds).
125For further explanation please see the internet-draft.
126.El
127.Pp
128Each congestion control module may also expose other MIB variables
129to control their behaviour.
130Note that both NewReno and CUBIC now support Hystart++ based on the version 3 of the internet-draft.
131.Sh Kernel Configuration
132All of the available congestion control modules may also be loaded
133via kernel configutation options.
134A kernel configuration is required to have at least one congestion control
135algorithm built into it via kernel option and a system default specified.
136Compilation of the kernel will fail if these two conditions are not met.
137.Sh Kernel Configuration Options
138The framework exposes the following kernel configuration options.
139.Bl -tag -width ".Va CC_NEWRENO"
140.It Va CC_NEWRENO
141This directive loads the NewReno congestion control algorithm.
142.It Va CC_CUBIC
143This directive loads the CUBIC congestion control algorithm and is included
144in GENERIC by default.
145.It Va CC_VEGAS
146This directive loads the vegas congestion control algorithm, note that
147this algorithm also requires the TCP_HHOOK option as well.
148.It Va CC_CDG
149This directive loads the cdg congestion control algorithm, note that
150this algorithm also requires the TCP_HHOOK option as well.
151.It Va CC_DCTCP
152This directive loads the dctcp congestion control algorithm.
153.It Va CC_HD
154This directive loads the hd congestion control algorithm, note that
155this algorithm also requires the TCP_HHOOK option as well.
156.It Va CC_CHD
157This directive loads the chd congestion control algorithm, note that
158this algorithm also requires the TCP_HHOOK option as well.
159.It Va CC_HTCP
160This directive loads the htcp congestion control algorithm.
161.It Va CC_DEFAULT
162This directive specifies the string that represents the name of the system default algorithm, the GENERIC kernel
163defaults this to CUBIC.
164.El
165.Sh SEE ALSO
166.Xr cc_cdg 4 ,
167.Xr cc_chd 4 ,
168.Xr cc_cubic 4 ,
169.Xr cc_dctcp 4 ,
170.Xr cc_hd 4 ,
171.Xr cc_htcp 4 ,
172.Xr cc_newreno 4 ,
173.Xr cc_vegas 4 ,
174.Xr tcp 4 ,
175.Xr config 5 ,
176.Xr config 8 ,
177.Xr mod_cc 9
178.Sh ACKNOWLEDGEMENTS
179Development and testing of this software were made possible in part by grants
180from the FreeBSD Foundation and Cisco University Research Program Fund at
181Community Foundation Silicon Valley.
182.Sh HISTORY
183The
184.Nm
185modular congestion control framework first appeared in
186.Fx 9.0 .
187.Pp
188The framework was first released in 2007 by James Healy and Lawrence Stewart
189whilst working on the NewTCP research project at Swinburne University of
190Technology's Centre for Advanced Internet Architectures, Melbourne, Australia,
191which was made possible in part by a grant from the Cisco University Research
192Program Fund at Community Foundation Silicon Valley.
193More details are available at:
194.Pp
195http://caia.swin.edu.au/urp/newtcp/
196.Sh AUTHORS
197.An -nosplit
198The
199.Nm
200facility was written by
201.An Lawrence Stewart Aq Mt lstewart@FreeBSD.org ,
202.An James Healy Aq Mt jimmy@deefa.com
203and
204.An David Hayes Aq Mt david.hayes@ieee.org .
205.Pp
206This manual page was written by
207.An David Hayes Aq Mt david.hayes@ieee.org
208and
209.An Lawrence Stewart Aq Mt lstewart@FreeBSD.org .
210