[954] in NetBSD-Development
Re: Dan Geer: PCI board rumor
daemon@ATHENA.MIT.EDU (Ken Raeburn)
Fri Aug 18 15:07:29 1995
Date: Fri, 18 Aug 1995 15:06:46 -0400
From: Ken Raeburn <raeburn@cygnus.com>
To: Dan Geer <geer@cam.ov.com>, netbsd-dev@MIT.EDU
In-Reply-To: Marc Horowitz's message of Fri, 18 Aug 1995 14:18:14 -0400
It appears that one of the common I/O controller chips used in PC
boards has a race condition that causes data corruption. This
controller is frequently used on board that use the PCI bus. One
estimate is that 1/3 of all the PCI PC's are affected. Systems that
have heavy I/O traffic (OS/2, Linux) are much more likely to see this
problem than DOS or Windows 3.1. Jury is still out on Windows 95.
Any info on what conditions trigger the problem and how to test for
it? Or chip ids so we can tell by inspection if a board is "bad"?
