From owner-freebsd-stable  Fri Mar 29 08:15:08 1996
Return-Path: owner-stable
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id IAA28633
          for stable-outgoing; Fri, 29 Mar 1996 08:15:08 -0800 (PST)
Received: from asstdc.scgt.oz.au (root@asstdc.scgt.oz.au [202.14.234.65])
          by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id IAA28610
          for <stable@freebsd.org>; Fri, 29 Mar 1996 08:14:52 -0800 (PST)
Received: (from imb@localhost)
          by asstdc.scgt.oz.au (8.7.5/BSD4.4) id DAA27473
          Sat, 30 Mar 1996 03:13:23 +1100 (EST)
From: michael butler <imb@scgt.oz.au>
Message-Id: <199603291613.DAA27473@asstdc.scgt.oz.au>
Subject: Re: Stable as bad with Adaptec as NCR
To: scrappy@ki.net (Marc G. Fournier)
Date: Sat, 30 Mar 1996 03:13:22 +1100 (EST)
Cc: stable@freebsd.org
In-Reply-To: <Pine.BSF.3.91.960329093949.568B-100000@ki.net> from "Marc G. Fournier" at Mar 29, 96 09:43:37 am
X-Mailer: ELM [version 2.4 PL24beta]
Content-Type: text
Sender: owner-stable@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Marc G. Fournier writes:

> 	Anyone care to make suggestions on maybe options I should turn on
> in the kernel?  This morning, less then 24hrs since last reboot, the system
> froze with a SCSI bus hang...again :( (hard drive LED on bright red)

 [ .. ]

> CPU: i486 DX4 (486-class CPU)
>   Origin = "GenuineIntel"  Id = 0x480  Stepping=0
>   Features=0x3<FPU,VME>
> real memory  = 16777216 (16384K bytes)
> avail memory = 14692352 (14348K bytes)
> Probing for devices on PCI bus 0:
> chip0 <SiS 85c496> rev 49 on pci0:5

This is a CPU with a write-back-capable L1 cache. Mine is an AMD without
this ability.

Some things I can think of are ..

i) check if your BIOS allows you to switch the L1 cache between "write-back"
and "write-thru" modes. If it's set to "write-back", try changing to
"write-thru". (L2 cache should already be set as "write-thru").

ii) if already "write-thru", try disabling the L1 cache entirely (:-(),

iii) make sure SCSI cable is OK and properly terminated,

iv) you have a reasonable number of drives on that box .. are they all on a
decent (!200W) supply or external ? If there's a shortage of connectors and
you're "splitting" rails, try to put a lightly loaded drive with a heavily
loaded one rather than two heavily loaded ones together. Garbage on SCSI bus
through supply shortfalls won't help.

v) make sure that other wait-state, bus clock and RAM burst-fetch parameters
are not more aggressive than mfr recommendation,

	michael