From owner-freebsd-current  Thu May  4 01:28:44 1995
Return-Path: current-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id BAA01864
          for current-outgoing; Thu, 4 May 1995 01:28:44 -0700
Received: from pelican.com (pelican.com [134.24.4.62])
          by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id BAA01858
          for <current@freebsd.org>; Thu, 4 May 1995 01:28:34 -0700
Received: by pelican.com (Smail3.1.28.1 #5)
	id m0s6wGp-000K0iC; Thu, 4 May 95 01:28 WET DST
Message-Id: <m0s6wGp-000K0iC@pelican.com>
From: pete@pelican.com (Pete Carah)
Subject: VLB VGA query
To: current@FreeBSD.org
Date: Thu, 4 May 1995 01:28:26 -0700 (PDT)
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 1394      
Sender: current-owner@FreeBSD.org
Precedence: bulk

In trying an install of the recent SNAP, I've run into a strange
problem.

System is an OPTI-based P90 with 3 VLB slots, 2 plain ISA (16bit)
slots, and 4 PCI slots.  I don't know the manufacturer.  Has an award
bios.

VLB IDE controller, Genoa (Cirrus accelerated) VLB VGA, 2 Conner 1280mb
IDE disks (which work fine and even configure and work in freebsd!!)

The problem is that there is data/attribute confusion in the vga, but
not under DOS at all.  only_ under freebsd (but I haven't tried Windoze...)
I could get through the install by knowing generally what was going on
and hitting ^L a lot during the fdisk/disklabel steps.  I presume that
I need to fix some delay somewhere...  Does anyone have a pointer to
what?

I've tried all the reasonable combinations of vlb and isa bus speed and
wait state jumpers (this one isn't soft-config'd for these speeds) with
no difference at all.  I've left it at 8.33 mhz isa and '<=33' vlb with
no vlb wait states (as the board came).

We don't need the vlb performance (this will be a server anyhow and
not run the X server) so will be substituting a cheap plain ISA vga card
when we get it, just to save the vlb one for someone that needs it.  I
assume this will help, but don't know.

However, has anyone seen this or something like it?
Are there known limitations of this chip set?  I could get chip numbers
and date codes if needed...

-- Pete