Subject: Re: configure() with valid context
To: Ignatios Souvatzis <ignatios@cs.uni-bonn.de>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: tech-kern
Date: 09/16/1999 09:22:24
On Thu, 16 Sep 1999 09:53:26 +0200 
 Ignatios Souvatzis <ignatios@cs.uni-bonn.de> wrote:

 > >   	initclocks();
 > > + 
 > > + 	/* Configure the system hardware.  This will enable interrupts. */
 > > + 	configure();
 > >   
 > >   #ifdef SYSVSHM
 > >   	/* Initialize System V style shared memory. */
 > > 
 > > ...essentially, this allows autoconfiguration to run in a valid thread
 > > context.
 > > 
 > > Objections?
 > 
 > I don't remember... where are the battery backed clocks read?
 > After autoconfiguration? 

Yes, actually, I realized there is a slight bug in my patch when I tried to
boot it on an Alpha :-)

I've since moved the initclocks() call into configure(), immediately after
cpu_configure() returns.  So, those semantics are now preserved.

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