>From: Peter Galbavy <peter@wonderland.org>
>Date: Tue, 18 Jun 1996 08:17:36 +0100 (BST)
>
>I am currently testing a very expensive (comparitively) rack mount
>PC from Texas Microsystems. It has a passive backplane, 6 PCI slots
>and 11 (!) ISA slots.

I've seen ads for this card, and I'm curious:  Do they violate the PCI
spec in order to get those six slots on there, or do they use
PCI-to-PCI bridges?

>Problem is, with -current, I am getting various user-land randomness,
>like core dumps, file not found etc etc. Occasionally I see:
>
>	NMI port 61 a0, port 70 ff

Port 61h is the KB controller's status port.  The MSB indicates RAM
parity errors; you have flaky RAM somewhere.  Flaky RAM -> flaky
system.  Turning off the cache will not fix the problem.

Port 70h is the RTC address/NMI enable port.  On standard RTC chips
it's write-only, so I don't see why NetBSD's NMI handler prints it.
Maybe it's readable on some systems.

<RANT>
Hey, PC industry, this is why we need parity-checked RAM!  Without it,
it's a lot harder to diagnose problems like Peter's.
</RANT>
-- 
Mike Long <mike.long@analog.com>     <URL:http://www.shore.net/~mikel>
VLSI Design Engineer         finger mikel@shore.net for PGP public key
Analog Devices, CPD Division          CCBF225E7D3F7ECB2C8F7ABB15D9BE7B
Norwood, MA 02062 USA       (eq (opinion 'ADI) (opinion 'mike)) -> nil