Path: utzoo!attcan!uunet!tut.cis.ohio-state.edu!zaphod.mps.ohio-state.edu!sdd.hp.com!hplabs!hpda!hpcupt1!swh
From: swh@hpcupt1.HP.COM (Steve Harrold)
Newsgroups: comp.sys.ibm.pc.hardware
Subject: Re: "I/O card parity interrupt at 156A:006E" ...
Message-ID: <51110002@hpcupt1.HP.COM>
Date: 17 Sep 90 15:40:10 GMT
References: <27889@pasteur.Berkeley.EDU>
Organization: Hewlett Packard, Cupertino
Lines: 25

>>> What does "I/O card parity interrupt at 156A: 006E" mean?
>>> "NMI (S) to shut off, (R) for Reboot, other key to continue",
>>> I have a 386 64k cache IBM compatible computer, phoenix BIOS 1.10.14,
>>> 4 meg RAM.
---------------
I have periodically experienced this with a 286 machine (HP Vectra) with
extra RAM, as well as with a 386 machine (also Vectra).  

The only way to reliably clear the problem is to power off, power on the
machine.  

I have a theory that the various managers used to control the above-1-meg
RAM are deficient and are not resetting the memory card correctly
under some conditions (e.g. when an application aborts the DOS kernel).
It seems that a system reset (ctl-alt-del) is not sufficient to reset the
memory board (perhaps a BIOS flaw?) and the alleged "I/O card parity.."
message manifests itself.

The card is definitely misbehaving, but it doesn't appear to be a hardware
problem.  The control software seems to be setting up the card for an
"impossible" scenario that the card was not designed to handle (nor was 
it required to do so).

I may be wrong (not being a hardware engineer) but perhaps my analysis will
prompt further comment.