xref: /freebsd/share/man/man4/h_ertt.4 (revision aca5021d5f7dcab1e11692923266373e35322d9a)
1.\"
2.\" Copyright (c) 2010-2011 The FreeBSD Foundation
3.\" All rights reserved.
4.\"
5.\" This documentation was written at the Centre for Advanced Internet
6.\" Architectures, Swinburne University of Technology, Melbourne, Australia by
7.\" David Hayes under sponsorship from the FreeBSD 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 January 18, 2012
33.Dt H_ERTT 4
34.Os
35.Sh NAME
36.Nm h_ertt
37.Nd Enhanced Round Trip Time Khelp module
38.Sh SYNOPSIS
39.In netinet/khelp/h_ertt.h
40.Sh DESCRIPTION
41The
42.Nm
43Khelp module works within the
44.Xr khelp 9
45framework to provide TCP with a per-connection, low noise estimate of the
46instantaneous RTT.
47The implementation attempts to be robust in the face of delayed
48acknowledgements, TCP Segmentation Offload (TSO), receivers who manipulate TCP
49timestamps and lack of the TCP timestamp option altogether.
50.Pp
51TCP receivers using delayed acknowledgements either acknowledge every second packet
52(reflecting the time stamp of the first) or use a timeout to trigger the
53acknowledgement if no second packet arrives.
54If the heuristic used by
55.Nm
56determines that the receiver is using delayed acknowledgements, it measures the
57RTT using the second packet (the one that triggers the acknowledgement).
58It does not measure the RTT if the acknowledgement is for the
59first packet, since it cannot be accurately determined.
60.Pp
61When TSO is in use,
62.Nm
63will momentarily disable TSO whilst marking a packet to use for a new
64measurement.
65The process has negligible impact on the connection.
66.Pp
67.Nm
68associates the following struct with each connection's TCP control block:
69.Bd -literal
70struct ertt {
71	TAILQ_HEAD(txseginfo_head, txseginfo) txsegi_q;	/* Private. */
72	long		bytes_tx_in_rtt;		/* Private. */
73	long		bytes_tx_in_marked_rtt;
74	unsigned long	marked_snd_cwnd;
75	int		rtt;
76	int		maxrtt;
77	int		minrtt;
78	int		dlyack_rx;			/* Private. */
79	int		timestamp_errors;		/* Private. */
80	int		markedpkt_rtt;			/* Private. */
81	uint32_t	flags;
82};
83.Ed
84.Pp
85The fields marked as private should not be manipulated by any code outside of
86the
87.Nm
88implementation.
89The non-private fields provide the following data:
90.Bl -tag -width ".Va bytes_tx_in_marked_rtt"  -offset indent
91.It Va bytes_tx_in_marked_rtt
92The number of bytes transmitted in the
93.Va markedpkt_rtt .
94.It Va marked_snd_cwnd
95The value of cwnd for the marked rtt measurement.
96.It Va rtt
97The most recent RTT measurement.
98.It Va maxrtt
99The longest RTT measurement that has been taken.
100.It Va minrtt
101The shortest RTT measurement that has been taken.
102.It Va flags
103The ERTT_NEW_MEASUREMENT flag will be set by the implementation when a new
104measurement is available.
105It is the responsibility of
106.Nm
107consumers to unset the flag if they wish to use it as a notification method for
108new measurements.
109.El
110.Sh SEE ALSO
111.Xr cc_chd 4 ,
112.Xr cc_hd 4 ,
113.Xr cc_vegas 4 ,
114.Xr mod_cc 4 ,
115.Xr hhook 9 ,
116.Xr khelp 9
117.Sh ACKNOWLEDGEMENTS
118Development and testing of this software were made possible in part by grants
119from the FreeBSD Foundation and Cisco University Research Program Fund at
120Community Foundation Silicon Valley.
121.Sh HISTORY
122The
123.Nm
124module first appeared in
125.Fx 9.0 .
126.Pp
127The module was first released in 2010 by David Hayes whilst working on the
128NewTCP research project at Swinburne University of Technology's Centre for
129Advanced Internet Architectures, Melbourne, Australia.
130More details are available at:
131.Pp
132http://caia.swin.edu.au/urp/newtcp/
133.Sh AUTHORS
134.An -nosplit
135The
136.Nm
137Khelp module and this manual page were written by
138.An David Hayes Aq david.hayes@ieee.org .
139.Sh BUGS
140The module maintains enhanced RTT estimates for all new TCP connections created
141after the time at which the module was loaded.
142It might be beneficial to see if it is possible to have the module only affect
143connections which actually care about ERTT estimates.
144