From owner-cvs-all  Tue Dec 17  9:57: 2 2002
Delivered-To: cvs-all@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP
	id 89B1A37B401; Tue, 17 Dec 2002 09:57:01 -0800 (PST)
Received: from repoman.freebsd.org (repoman.freebsd.org [216.136.204.115])
	by mx1.FreeBSD.org (Postfix) with ESMTP
	id 4AF2643ED1; Tue, 17 Dec 2002 09:57:01 -0800 (PST)
	(envelope-from sam@FreeBSD.org)
Received: from repoman.freebsd.org (localhost [127.0.0.1])
	by repoman.freebsd.org (8.12.6/8.12.6) with ESMTP id gBHHv1mV028420;
	Tue, 17 Dec 2002 09:57:01 -0800 (PST)
	(envelope-from sam@repoman.freebsd.org)
Received: (from sam@localhost)
	by repoman.freebsd.org (8.12.6/8.12.6/Submit) id gBHHv1g0028419;
	Tue, 17 Dec 2002 09:57:01 -0800 (PST)
Message-Id: <200212171757.gBHHv1g0028419@repoman.freebsd.org>
From: Sam Leffler <sam@FreeBSD.org>
Date: Tue, 17 Dec 2002 09:57:01 -0800 (PST)
To: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org
Subject: cvs commit: src/sys/dev/ubsec ubsec.c ubsecvar.h
X-FreeBSD-CVS-Branch: RELENG_4
Sender: owner-cvs-all@FreeBSD.ORG
Precedence: bulk
List-ID: <cvs-all.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo@FreeBSD.ORG?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo@FreeBSD.ORG?subject=subscribe%20cvs-all>
List-Unsubscribe: <mailto:majordomo@FreeBSD.ORG?subject=unsubscribe%20cvs-all>
X-Loop: FreeBSD.ORG

sam         2002/12/17 09:57:01 PST

  Modified files:        (Branch: RELENG_4)
    sys/dev/ubsec        ubsec.c ubsecvar.h 
  Log:
  o split hst_invalid statistic into separate buckets to help diagnose problems
  o add statistics to track the queue length on the host and on the chip and
    count the number of times the h/w reports it's too busy to accept more
    requests
  o change default batching parameter from 2 to 1 (no batching) for optimal
    performance but at the cost of additional interrupts
  o add a ubsec_maxaggr parameter that controls the number of mcr1 ops passed
    to the h/w and default it to 1 for optimal performance; setting this value
    to >1 will (potentially) increase interrupt coalescing but at the cost of
    increased latency since the host only gets an interrupt when the entire batch
    of requests is completed (tests show this is bad for network performance)
  
  Sponsored by:   Vernier Networks
  
  Revision  Changes    Path
  1.6.2.3   +54 -11    src/sys/dev/ubsec/ubsec.c
  1.2.2.2   +9 -1      src/sys/dev/ubsec/ubsecvar.h

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message