SCO Xenix/386 and large drives

[Chip Rosenthal]

In article <80 at iisat.UUCP> kevin at iisat.UUCP (Kevin Davies) writes:
>Does anyone know whether or not SCO Xenix/386 2.2.2 can handle
>hard drives that have more than 1024 cylinders (under ESDI) ?
>Is an upgrade to this version required to hande a drive this large ?
>Any help or hints would be greatly appreciated.

I just spent a few weeks researching a 386 XENIX system with RLL drives.
Some of what I found out (much of it thanks to you netfolk out there) is
probably applicable to this.

There are two approaches.  First, you can run the OMTI8627 ESDI/RLL
controller with a special version of XENIX.  (I believe that there is a
second ESDI/MFM controller.)  This version, a.k.a. the "ESDI version",
replaces the standard hard disk driver with one which talks to the 8627.
There is about a $100 premium for this version.  I strongly recommend
against this approach.  First, the 8627 is relatively expensive and slow.
Second, it ain't made anymore.  I'd be real concerned about if and how
SCO will continue to support an obsolete piece of hardware like this.
And if you want 2.3 and 3.0 -- which you should -- you need to be concerned
about the upgrade path.

The second approach is to get a controller which looks like a standard
ST506 controller, but in fact talks to ESDI disks.  In this case, you can
run with the standard 386 XENIX release.  My understanding is that the
Western Digital WD1007 and Adaptec 2322 do this.

The following information is quoted from SCO's 2.2.2 Release Notes (Apr 88):

  Many hard disks, both standard and nonstandard, can be used by the XENIX
  System V Operating System as long as the disk controller supports the
  drive.  The disk must interface electrically with the disk controller
  (usually "ST506").  Controllers supporting other drive interfaces such
  as RLL or ESDI will work as long as the controller presents an
  ST506-compatible interface to the bus.

The section goes on to say that either the motherboard ROM must contain
an entry describing the disk (e.g. heads, cylinders, sectors, etc.), or
you can use "dkinit" to define a non-standard set of parameters.  So,
you should be able to enter the information on a large disk even if there
isn't a BIOS entry for it.

It appears that if you can find a controller which looks like a WD1003
*and* it requires no funky support routines (e.g. strange .SYS files are
required to use it with DOS), then it will work.
-- 
Chip Rosenthal /// chip at vector.UUCP /// Dallas Semiconductor /// 214-450-0400
{uunet!warble,sun!texsun!rpp386,killer}!vector!chip
I won't sing for politicians.  Ain't singing for Spuds.  This note's for you.