Subject: Re: machine check on 1.4
To: Nathan J. Williams <nathanw@MIT.EDU>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: port-alpha
Date: 09/25/1999 12:02:58
On 25 Sep 1999 14:58:46 -0400 
 nathanw@MIT.EDU (Nathan J. Williams) wrote:

 > The backtrace wouldn't be very interesting; the panic is triggered by
 > the asynchronous machine check interrupt. I'm told that all the
 > information needed to interpret the machine check is in the
 > mces/vector/param data, but the actual interpretation is some kind of
 > Digital Unix^WTru64 state secret. 

that's pretty much the case... param points to something called a
"logout frame", which has CPU, core logic, and system variant specific
data in it.

...however, you'll notice that NetBSD can interpret the logout frame on
some systems... i.e. some of us have the documentation necessary to
write this code... it's just a matter of doing it, really.

        -- Jason R. Thorpe <thorpej@nas.nasa.gov>