From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Fri, 16 Feb 1996 18:56:19 -0500
From: jim mckie jmk@plan9.att.com
Subject: Problems with ide drive on a very old machine
Topicbox-Message-UUID: 3b72f536-eac8-11e9-9e20-41e7f4b1d025
Message-ID: <19960216235619.eAQd_HJRY5QDcaqem4gxYSnF6p5JGlR37PL_4nsw0ps@z>

It's likely that the drive is slow in responding to an Ident message,
you can try turning on the debug trace in devata.c (DPRINT) to see where
things are sticking. It may be some of the timeouts need to be adjusted,
they are from the ATA 'specification' and don't always work well with
older hardware. If you turn tracing on, it's best to boot a kernel
over the ether and to not automatically start up anything that requires
the disc. Then "ls '#H'" will give a trace.

------ forwarded message follows ------

>>From cse.psu.edu!9fans-outgoing-owner Fri Feb 16 18:23:05 EST 1996
Received: from colossus.cse.psu.edu by plan9; Fri Feb 16 18:23:05 EST 1996
Received: by colossus.cse.psu.edu id <78511>; Fri, 16 Feb 1996 18:12:55 -0500
Received: from up8.univ-paris8.fr ([193.54.155.1]) by colossus.cse.psu.edu with SMTP id <78696>; Fri, 16 Feb 1996 18:08:10 -0500
Received: from tao.ai.univ-paris8.fr (tao.ai.univ-paris8.fr [192.33.156.85]) by up8.univ-paris8.fr (8.6.11/8.6.9) with SMTP id XAA23336 for <9fans@cse.psu.edu>; Fri, 16 Feb 1996 23:53:53 +0100
Received: from ching.ai.univ-paris8.fr by tao.ai.univ-paris8.fr (4.1/SMI-4.1)
	id AA20813; Sat, 17 Feb 96 00:02:00 +0100
Date:	Fri, 16 Feb 1996 18:02:00 -0500
Message-Id: <9602162302.AA20813@tao.ai.univ-paris8.fr>
Received: by ching.ai.univ-paris8.fr (4.1/SMI-4.1)
	id AA01212; Sat, 17 Feb 96 00:00:40 +0100
From:	Jean Mehat <tao.ai.univ-paris8.fr!jm>
To:	cse.psu.edu!9fans
Subject: Problems with ide drive on a very old machine
Sender: cse.psu.edu!owner-9fans
Precedence: bulk
Reply-To: cse.psu.edu!9fans

I am trying to use as an authentification server an old 386sx16. With the
kernel installed by the installation procedure, the disk hanged and could
not be accessed before a reset (but the machine booted via il).

After applying the updates (and the forsyth's patch found in 9fans to
unlock the keyboard), I can boot 9pcdisk on the local disk if I set
the root on the file server. At boot time, I have:
	session...fs...kfs init 6: can't read #H/hd0fs
and the boot proceeds. Strangely enough, the hd0* files appear in /dev.
Once I have a rc running:
	term$ disk/kfs -f /dev/hd0fs
	kfs init 67: can't read /dev/hd0fs
	term$ disk/kfs -f /dev/hd0fs
	term$ mount /srv/kfs /mnt
	term$ cd /mnt; cat readme
I.e: the first kfs seems to fail, the second succeeds.
Before plunging in devata.c, has anyone an idea? Does the problem
most likely comes from the controller or from the disk?