[comp.unix.sysv386] <SUMMARY> ISC vs Adaptec 1542b vs 80387

phillip@BARTAL.COM (Phillip M. Vogel) (05/06/91)

In article <451@bartal.BARTAL.COM> I lamented:

>Due to a change in system requirements at bartal, I decided to
>remove the 80387 from the unix box, and put it in the dos box
>where I run most of my CAD applications.  Well, The unix machine
>won't get past "loading the unix system" with the 387 removed.
>However, if I also pull the SCSI controller, it will boot just
>fine from the st-4096.  I checked the cmos, and all is correct,
>as far as I can tell.

Well, thanks to all the people who responded.  It turns out that the
problem was pretty straightforward, but there were several things
wrong at the same time (just to keep things confusing).

1. There are jumpers on the board that have to be removed to
reflect the lack of coprocessor (even though the ROM setup
program shows that the 387 is gone with the jumpers still in).

2. ISC uses different CMOS bit(s) to reflect the type and
presence of a coprocceser than Phoenix does.  I plum forgot that
I had had this problem "way back when", and had gotten a program
from ISC to manually set the CMOS (which I had to unset).

3. The damned security bug!  When this became an issue some
months ago, I, having a coprocessor installed, simply disabled
the FP emulator by setting UAREAUS and UAREARW to 0 and
kconfigging (new word).  This was fine, but if you remove the
387, you can't do any floating point.

Having fixed all of this, bartal is now 387-less and SCSIfied,
and I'm happy (and humbled). 
	Phillip
--
Phillip M. Vogel, President             | #include "/disclaimers/std.h"
Bartal Design Group, Inc.               | Domain: phillip@bartal.com
318 Marlboro Road, Englewood, NJ 07631  | (201)567-1343   FAX:(201)568-2891