On Wed, 12 Jan 2000, Benjamin Herrenschmidt wrote:

> On Tue, Jan 11, 2000, Bill Studenmund <wrstuden@nas.nasa.gov> wrote:
> 
> >Well, the problem is that we're also supposed to shut things down when
> >we're running, by quiescing OF. As I understand it, THAT requirement
> >shoots down nvram access and beeping. :-(
> 
> I theory, RTAS is not killed by quiesce since it's meant to be used by
> the host OS after the quiesce call. In practice, RTAS doesn't work. So
> you'll have to write direct PCI config access routines and NVRAM access.

Direct PCI config access itself is not a problem. We'd need to know where
the windows are, but our bus space code was designed with it in mind so
all that has to happen is that the macppc code take advantage.

> On the Uni-N machines, NVRAM is actually a flash, I'm currently trying to
> figure out the flash algorithm. 

That would be cool to know! Good luck!

Take care,

Bill