Subject: Re: Unexpected problems, part 2
To: None <blaz.antonic@siol.net>
From: James K. Lowden <jklowden@schemamania.org>
List: port-vax
Date: 05/20/2003 04:15:08
On Tue, 20 May 2003 08:35:28 -0700, Blaz Antonic <blaz.antonic@siol.net>
wrote:
> > > Another question i thought of, at the risk of sounding silly:
> > > is this break to console thing handled by some interrupt or
> > > something ? Shouldn't OS be able to hook that interrupt and
> > > prevent this behavior if that is the case ?
> > 
> > It almost certainly is an NMI of some sort but there will
> > be no easy way for the OS to get in the way. The whole point
> > is to be able to still recover the console even when something
> > else decides to play at IPL 31 forever!
> 
> I don't know how things work on Vaxen but NMIs can be hooked too on x86
> so in theory the handler could decide whether it wants to invoke
> original handler (the one that breaks to console and so on) or not.
> Actually i'm slightly surprised that OS allows such interruptions,
> unless, of course, if this is a 'feature' :)

Check.  Feature, yes.  

You're probably used to the PC world, where the distinction between
hardware and software isn't so clear, and where the firmware is basically
not capable of much.  PC's don't have a 'halt' button; access to the
firmware (BIOS) is available only at boot time.  

On the VAX, the firmware is always available.  Halt the system, diddle
with the firmware (say, check or change the auto-reboot option), and
continue.  While you're there, the OS is "on hold".  No muss no fuss.  

--jkl