On Sun, 15 Dec 2002, Martin Husemann wrote:

> On Sat, Dec 14, 2002 at 11:47:07PM -0800, Bill Studenmund wrote:
>
> > I think we can get what everyone wants with something more of an
> > intermediate step. Here's its proposal:
>
> To play devils advocate: when we change disklabels, effectively inventing
> a new format for NetBSD-native storage of this info (in all other cases we
> don't have a choice, i.e. a sun will have to have a sun compatible disklabel
> to load ofwboot at least) - why don't we go all the way and use this strange
> new standard format that Intel enforces for its 64bit stuff - sorry I can't
> remember the acronym.
>
> It has well defined interaction with MBR, allows 16k slices, is completely
> new and incompatible as well but is a standard (tm).

Let's forget the notion of a NetBSD-native disklabel. One-size-fits-all is
what got us into this mess.

I think that you're right though that whatever the Intel thing is, we
should support it, and make sure what we do here will work w/ it.

> ... or, of course, just go to wedges, only providing enough info for the first
> parts of the boot process in (firmware-) native format.

What about all those questions I raised about wedges? :-)

Take care,

Bill