On Mon, 8 Nov 1999, Bob Nestor wrote:

> I was working with Lennart yesterday on this. He sent me some patches and 
> a new file.  We got to the hang point and decided that it's probably some 
> kind of interrupt problem.  They seem to be blocked or totally disabled.  
> He was going to do some more digging on the problem.
> 
> The 1.4.1 kernel doesn't seem to have this problem since that's what I 
> was able to run to do my initial filesystem installation.  I don't know 
> how much has changed since then, but somewhere in those changes there has 
> to be a clue or two.

I suspect this is the problem. I wonder if we're dropping into OF at any
point in here...

My initial thought was that we might want to wrap calls to OF with a form
of splfoo()/splx() pair. But what I'm thinking of would be a little
different in that we'd be saving the state of MSR[EE] (sp?) - the
interrupt enable bit (External Exception enable) - and restore it
afterwards. The current spl routines, AFAICT, only twiddle internal
interrupt masks.

Take care,

Bill