From owner-freebsd-current  Mon Dec 13  0:17:39 1999
Delivered-To: freebsd-current@freebsd.org
Received: from freebsd.dk (freebsd.dk [212.242.42.178])
	by hub.freebsd.org (Postfix) with ESMTP
	id 5794814D07; Mon, 13 Dec 1999 00:17:36 -0800 (PST)
	(envelope-from sos@freebsd.dk)
Received: (from sos@localhost)
	by freebsd.dk (8.9.3/8.9.1) id JAA37135;
	Mon, 13 Dec 1999 09:17:35 +0100 (CET)
	(envelope-from sos)
From: Soren Schmidt <sos@freebsd.dk>
Message-Id: <199912130817.JAA37135@freebsd.dk>
Subject: Re: ATA weird message?
In-Reply-To: <199912130739.XAA05787@mass.cdrom.com> from Mike Smith at "Dec 12, 1999 11:39:59 pm"
To: msmith@FreeBSD.ORG (Mike Smith)
Date: Mon, 13 Dec 1999 09:17:35 +0100 (CET)
Cc: haro@tk.kubota.co.jp (Munehiro Matsuda),
	freebsd-current@FreeBSD.ORG
X-Mailer: ELM [version 2.4ME+ PL54 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
Sender: owner-freebsd-current@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

It seems Mike Smith wrote:
> > It seems Munehiro Matsuda wrote:
> > > Hi all,
> > > 
> > > I am using -current as of December 9 (CTM:src-cur.4130.gz), and
> > > got following weird ATA related messages while 'make -j4 buildworld'.
> > > I never had this kind of message when using wd drivers.
> > > 
> > > 	ata0-master: ad_timeout: lost disk contact - resetting
> > > 	ata0: resetting devices .. done
> > 
> > Hmm, maybe the timeout in ata-disk.c is too short, try increasing
> > the 5*hz to say 10*hz in line 436, and see if that changes anything..
> 
> Ugh.  Are we still using these short timeouts for I/O transaction 
> completion?  Much pain some time back established that we need at least 
> 30 seconds for some drives doing internal error recovery, and as long as 
> the drive is returning something sensible (ie. still busy), we should 
> give it at least that long.

Not really, the timeout for completing a request on a drive that
said it is ready, _should_ complete in much less than 5 secs, but
a reset of a hung drive should be waited for 31 secs, which we
do allready in ata_reset()...

I'm afraid that the sporious timeouts that some are seeing is an
artifact of another problem...

-Søren


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