Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!ub!dsinc!bagate!cbmvax!andy
From: andy@cbmvax.commodore.com (Andy Finkel)
Newsgroups: comp.sys.amiga
Subject: Re: 2091 Interface
Keywords: In article <5357@crash.cts.com> denny@pnet01.cts.com (Dennis Anderson) writes:
Message-ID: <15515@cbmvax.commodore.com>
Date: 1 Nov 90 15:16:45 GMT
References: <5357@crash.cts.com> <13925@mcdphx.phx.mcd.mot.com>
Reply-To: andy@cbmvax.commodore.com (Andy Finkel)
Organization: Commodore, West Chester, PA
Lines: 55

In article <13925@mcdphx.phx.mcd.mot.com> teroach!stan@phx.mcd.mot.com (Stan Fisher) writes:
>I know many of us 2091 owners (590 too) have experienced this problem..  I
>asked a while back on the net about the SCSI bus hang problem (multiple
>drives on the buss going poof) and was told to disable the reselection
>option in the HDTOOLBOX program's advanced options area.  This was
>"supposed to fix the problem" which in reality is only a patch workaround
>to reduce the occurance of the problem.  

Since the cause of the data loss is an extra interrupt during the SCSI
reselect, turning off reselection on all drives will reduce the
occurance of the problem to zero.

>The REAL problem is that Commodore
>got a batch of bad Western Digital SCSI controller chips with reselection
>problems and is trying to avoid doing the right thing and fixing everyones
>boards with new controller chips! I also was told that the firmware

You are incorrect.  The problem is that Western Digital came out with
a *new* revision of their chip that works slightly differently than
the old one; its only the new chips that show the problem.  Their
low level SCSI state machine has changed; since we may be the only
ones using their low level command set, and the only ones who
are using (or want to use) reselection on disk drives, 
(as well as supporting multitasking so you could initiate two copies)
we got bit by the new chip.

>me back with the answer.  About 1 week later I did get a call.  This person
>said another rev. of the firmware was just about ready, and when it got to
>the dealers, it would be offered as a free upgrade.  I wanted to know how
>they planned to fix the hardware problem with software. He wasn't sure but he
>said "they hoped it wouldn't effect the performance of the controller"

This is one of the primary functions of software.  As en employee of a chip
company, you should not be surprised.  The software is the last thing
to get done....the chips are cast in stone long, long before the software
is done.  (for instance, the 030 cache requires a software kludge to work
properly)

> !!!!  WHAT !!!!  Gee I hope so too!!!  If my company worked things this
>way you can bet we'd be losing alot of business.  How would CBM like it if

And, sometimes you take a performance hit...for instance, in the single
drive case, you get higher numbers if you don't support reselect at all;
supporting it fully is a hit, no matter what you do.

			andy
-- 
andy finkel		{uunet|rutgers|amiga}!cbmvax!andy
Commodore-Amiga, Inc.

"It is much easier to suggest solutions when you know nothing about the
 problem."

Any expressed opinions are mine; but feel free to share.
I disclaim all responsibilities, all shapes, all sizes, all colors.