From owner-freebsd-stable  Tue May 26 17:44:35 1998
Return-Path: <owner-freebsd-stable@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id RAA00358
          for freebsd-stable-outgoing; Tue, 26 May 1998 17:44:35 -0700 (PDT)
          (envelope-from owner-freebsd-stable@FreeBSD.ORG)
Received: from public.bta.net.cn (public.bta.net.cn [202.96.0.97])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id RAA00340
          for <freebsd-stable@FreeBSD.ORG>; Tue, 26 May 1998 17:44:26 -0700 (PDT)
          (envelope-from robinson@public.bta.net.cn)
Received: (from robinson@localhost)
	by public.bta.net.cn (8.8.5/8.8.5) id IAA12575;
	Wed, 27 May 1998 08:42:55 +0800 (GMT)
Date: Wed, 27 May 1998 08:42:55 +0800 (GMT)
From: Michael Robinson <robinson@public.bta.net.cn>
Message-Id: <199805270042.IAA12575@public.bta.net.cn>
To: freebsd-stable@FreeBSD.ORG, haskin@ptway.com
Subject: Re: Bug in wd driver
Sender: owner-freebsd-stable@FreeBSD.ORG
Precedence: bulk


"Brian Haskin" <haskin@ptway.com> writes:
>I think what Mike is saying is that at least in this case the bad spot
>doesn't seem to be the real problem. The problem appears to be that the disk
>firmware is crashing, which is certainly an unrecoverable error.

He may be saying this, but, quoting from my original message:

<quote>
 2. After this, the process requesting the I/O will be completely
    locked, but the disk will continue to make seeking noises 
    continuously until the system is powered off.
    Other processes are able to access the affected slice/partition
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    (ls, cat, etc.) without any difficulty, as long as they avoid 
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    the 7 affected sectors.  Any process which requires privileged
    ^^^^^^^^^^^^^^^^^^^^^^^
    kernel calls (halt, ps, etc.) will lock immediately and 
    completely.
</quote>

The disk firmware absolutely did not crash.  Nor, in fact, did the 
disk controller, nor even the driver itself.  Not even the kernel
(mostly).  No crashes.  A queued I/O request that never timed out, yes.
Crash, no.

	-Michael Robinson


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