31 Mar
2008
31 Mar
'08
5:25 p.m.
On 29/03/08 02:26 +0100, Peter Stuge wrote:
On Fri, Mar 28, 2008 at 08:42:41AM -0600, Jordan
Crouse wrote:
I prefer to educate users rather than bending around a bad habit.
The problem is - our users aren't interested in being educated, they are
interested in making their stuff work. Taking the high road benefits
nobody (yeah, I know that this is the story of the legacy PC world, but
its the world we live in).
I vote
for nvram, too, especially if we plan going to non-x86
architectures again. CMOS is really an odd term, I think it got
molded by the legacy bios vendors over the years.
On the other hand, CMOS is the agreed upon term for x86 platforms -
it sounds stupid to our ears, but we're not typical users. I would
stick with CMOS. I don't agree at all that coreboot is targeted
exclusively at x86.
If other architectures can use new bootcode like the PC then I would
love for coreboot to be that new bootcode.
Yes, today it is all x86, but if the hardware can be used in other
architectures then the software should as well.
In a few desktop machine generations, when Windows has been better
abstracted, the PC arch will linger no more.
AMD/NatSemi already threw it out with the Geode arch.
I think we'll have better code in v4 if we already keep other archs
in mind. (Not jump through hoops to befriend them, but just keep them
in mind.)
This won't come as a surprise to anybody, but we are lagging years behind
the other architectures. There are already very good free and open
source boot monitors for many of the other architectures. I view the
coreboot effort not as creating something new and magical, but rather
as dragging the x86 into modern times to catchup with its cousins that
long ago passed it in terms of cluefulness. I hate the idea of spending
valuable cycles thinking about other architectures when we're still so
far away from making our primary one work. Just my opinion.
Jordan
--
Jordan Crouse
Systems Software Development Engineer
Advanced Micro Devices, Inc.