Provide new socket option TCP_CCALGOOPT, which stands for TCP congestion

control algorithm options.  The argument is variable length and is opaque
to TCP, forwarded directly to the algorithm's ctl_output method.

Provide new includes directory netinet/cc, where algorithm specific
headers can be installed.

The new API doesn't yet have any in tree consumers.

The original code written by lstewart.
Reviewed by:	rrs, emax
Sponsored by:	Netflix
Differential Revision:	https://reviews.freebsd.org/D711

3 files changed, 36 insertions, 5 deletions

diff --git a/share/man/man4/mod_cc.4 b/share/man/man4/mod_cc.4
index f5f4493..4712a39 100644
--- a/share/man/man4/mod_cc.4
+++ b/share/man/man4/mod_cc.4
@@ -30,7 +30,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd January 12, 2015
+.Dd January 21, 2016
 .Dt MOD_CC 4
 .Os
 .Sh NAME
@@ -49,7 +49,9 @@ using the
 facility.
 .Pp
 The default algorithm is NewReno, and all connections use the default unless
-explicitly overridden using the TCP_CONGESTION socket option (see
+explicitly overridden using the
+.Dv TCP_CONGESTION
+socket option (see
 .Xr tcp 4
 for details).
 The default can be changed using a
@@ -57,6 +59,14 @@ The default can be changed using a
 MIB variable detailed in the
 .Sx MIB Variables
 section below.
+.Pp
+Algorithm specific parameters can be set or queried using the
+.Dv TCP_CCALGOOPT
+socket option (see
+.Xr tcp 4
+for details).
+Callers must pass a pointer to an algorithm specific data, and specify
+its size.
 .Sh MIB Variables
 The framework exposes the following variables in the
 .Va net.inet.tcp.cc
diff --git a/share/man/man4/tcp.4 b/share/man/man4/tcp.4
index 8c5887f..de993e7 100644
--- a/share/man/man4/tcp.4
+++ b/share/man/man4/tcp.4
@@ -34,7 +34,7 @@
 .\"     From: @(#)tcp.4	8.1 (Berkeley) 6/5/93
 .\" $FreeBSD$
 .\"
-.Dd October 27, 2015
+.Dd January 21, 2016
 .Dt TCP 4
 .Os
 .Sh NAME
@@ -137,6 +137,11 @@ send window size,
 receive window size,
 and
 bandwidth-controlled window space.
+.It Dv TCP_CCALGOOPT
+Set or query congestion control algorithm specific parameters.
+See
+.Xr mod_cc 4
+for details.
 .It Dv TCP_CONGESTION
 Select or query the congestion control algorithm that TCP will use for the
 connection.
diff --git a/share/man/man9/mod_cc.9 b/share/man/man9/mod_cc.9
index f1cd5be..05205ed 100644
--- a/share/man/man9/mod_cc.9
+++ b/share/man/man9/mod_cc.9
@@ -31,7 +31,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd December 26, 2014
+.Dd January 21, 2016
 .Dt MOD_CC 9
 .Os
 .Sh NAME
@@ -40,7 +40,8 @@
 .Nm CCV
 .Nd Modular Congestion Control
 .Sh SYNOPSIS
-.In netinet/cc.h
+.In netinet/tcp.h
+.In netinet/tcp_cc.h
 .In netinet/cc/cc_module.h
 .Fn DECLARE_CC_MODULE "ccname" "ccalgo"
 .Fn CCV "ccv" "what"
@@ -74,6 +75,7 @@ struct cc_algo {
 	void	(*cong_signal) (struct cc_var *ccv, uint32_t type);
 	void	(*post_recovery) (struct cc_var *ccv);
 	void	(*after_idle) (struct cc_var *ccv);
+	int	(*ctl_output)(struct cc_var *, struct sockopt *, void *);
 };
 .Ed
 .Pp
@@ -166,6 +168,20 @@ function is called when data transfer resumes after an idle period.
 It should be implemented to adjust state as required.
 .Pp
 The
+.Va ctl_output
+function is called when
+.Xr getsockopt 2
+or
+.Xr setsockopt 2
+is called on a
+.Xr tcp 4
+socket with the
+.Va struct sockopt
+pointer forwarded unmodified from the TCP control, and a
+.Va void *
+pointer to algorithm specific argument.
+.Pp
+The
 .Fn DECLARE_CC_MODULE
 macro provides a convenient wrapper around the
 .Xr DECLARE_MODULE 9