Newsgroups: comp.os.msdos.desqview
Path: utzoo!utgpu!watserv1!maytag!xenitec!geoffw
From: geoffw@xenitec.on.ca (Geoffrey Welsh)
Subject: Re: Perstor ARLL controller and DESQview
Organization: Xenitec Consulting Services, Kitchener, ON, CANADA
Date: Sat, 01 Jun 1991 18:14:36 GMT
Message-ID: <1991Jun01.181436.1833@xenitec.on.ca>
Summary: It's Pertor's fault.
References: <31MAY91094739@vms.huji.ac.il>

(BTW, thanks for the note on LOADHI's /GS parameter!)

In article <31MAY91094739@vms.huji.ac.il> gilf@vms.huji.ac.il writes:
>However, since I installed this controller, DESQview hang, in almost
>every session, and communications are interupted. The dealer I baught
>the controller from refuses to acknowlage the problem, puting is down
>to the software (what else).

   No, the fault lies almost exclusively in the Perstor's firmware.

>Does anyone else have any information? I think the problem is that the 
>Perstor controller 1) uses memory locations that conflict with DESQview,

   I doubt it; I don't know why your machine is hanging, but
losing characters on serial I/O is a common problem with those
Perstor controllers, even without DesqView.  The normal solution
is to invest in an NS16550AF chip, which will give the CPU much
longer to fetch the incoming characters.

>and, 2) uses an interupt that conflicts with communication.

   As far as I know, there is no interrupt conflict, but the
Perstor card either masks the interrupts or engages in very
long DMA bursts, either of which will prevent the CPU from
reading the incoming characters before they are lost.

>Does anyone have any idea if those problems can be resolved using software
>alone? Can anyone send me the address for Perstor (the one on the manual
>is either wrong, or outdated)?

   As I said, I don't know if there's a quick solution to your
machine hanging, but a 16550 will go far to help your serial
I/O problems.

   Geoff

