tron@zhadum.de (Matthias Scheler) writes:

> In article <mtud5t3vllv.fsf@contents-vnder-pressvre.mit.edu>,
> 	"Nathan J. Williams" <nathanw@wasabisystems.com> writes:
> > What I think would be genuine forward progress would be to use ACPI ...
> 
> I agree that ACPI is the better way to do that. But from reading various
> articles about getting Linux to run on notebooks it seems that there
> are still a lot of BIOS implementations with buggy ACPI support. So
> being able to run a non-ACPI kernel on such a system is a big advantage.

Hmm. ACPI can be used for lots of things... I would expect the
device-identifying part of it to be pretty solid, even if the
power-management stuff is dodgier.

> > ... to attach the remaining legacy devices and punt the ISA infrastructure
> > entirely.
> 
> There are still systems with real ISA buses e.g. P4 motherboards with
> a few ISA slots(*). While it is probably ok not to support them in the
> default kernels it would be good to allow ISA if somebody needs it.

Are there EM64T-capable motherboards with ISA? Vanilla P4 isn't so
interesting for the amd64 port...

        - Nathan