From owner-cvs-sys  Fri Mar 22 10:09:29 1996
Return-Path: owner-cvs-sys
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id KAA18680
          for cvs-sys-outgoing; Fri, 22 Mar 1996 10:09:29 -0800 (PST)
Received: (from wollman@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id KAA18655
          Fri, 22 Mar 1996 10:09:23 -0800 (PST)
Date: Fri, 22 Mar 1996 10:09:23 -0800 (PST)
From: Garrett Wollman <wollman>
Message-Id: <199603221809.KAA18655@freefall.freebsd.org>
To: CVS-committers, cvs-all, cvs-sys
Subject: cvs commit:  src/sys/netinet tcp_input.c tcp_var.h
Sender: owner-cvs-sys@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

wollman     96/03/22 10:09:22

  Modified:    sys/netinet  tcp_input.c tcp_var.h
  Log:
  A number of performance-reducing flaws fixed based on comments
  from Larry Peterson &co. at Arizona:
  
  - Header prediction for ACKs did not exclude Fast Retransmit/Recovery.
  - srtt calculation tended to get ``stuck'' and could never decrease
    when below 8.  It still can't, but the scaling factors are adjusted
    so that this artifact does not cause as bad an effect on the RTO
    value as it used to.
  
  The paper also points out the incr/8 error that has been long since fixed,
  and the problems with ACKing frequency resulting from the use of options
  which I suspect to be fixed already as well (as part of the T/TCP work).
  
  Obtained from:	Brakmo & Peterson, ``Performance Problems in BSD4.4 TCP''
  
  Revision  Changes    Path
  1.39      +54 -3     src/sys/netinet/tcp_input.c
  1.31      +23 -1     src/sys/netinet/tcp_var.h