Re: dev/ofw/ofdisk.c

[Tim Rightnour <root%garbled.net@localhost>]

On 16-Oct-2007 Frank Wille wrote:
> Hm. I noticed that you also removed the 'platform'-stuff in ofppc. So your
> recommended method woud be to start a completely new port for each platform?
> For example a pegasos-port, in parallel to macppc and prep?

No, not at all.  I'm simply not a fan of how the platform stuff was done, and
I'm not convinced most of it would translate over to the new order very well.

I simply don't think the differences in these platforms is quite so vast, once
you start using the hardware instead of the OFW for everything.  On prep, we
have a rather diverse set of machines that we work on, but by using the
detection routines, we can just find those differences that actually matter,
and deal with those on a one-by-one basis.  In my opinion, if a machine is so
fundamentally different than all the others, that it requires pretty much every
routine in the kernel to be a pointer, it might as well be a different port.

However, I think we will find that as we explore more of these ofppc ports,
that that is unlikely to occur.  Macppc is a rather good example however, as it
is so foreign to the ofppc model, that it really is it's own port.

I want to start fresh, and only create these differences when we find that they
are absolutely necc.

---
Tim Rightnour <root%garbled.net@localhost>
NetBSD: Free multi-architecture OS http://www.netbsd.org/
Genecys: Open Source 3D MMORPG: http://www.genecys.org/