Ronald G Minnich wrote:
>
> but consider this. We want to boot linux, openbsd, freebsd, netbsd, and
> maybe even Plan 9. All those systems have very different ways of doing
> things -- particularly for SMP. What happens when an OS calls linuxbios?
> how do we manage SMP? what's the right way to handle an interrupt while
> we're in the BIOS? Just locking out the OS for the time we're in the BIOS
> seems like a bad idea. What if an OS calls the BIOS, and "something bad"
> happens, and the BIOS can't get back to the OS? This happens all the time
> on my Thinkpad -- the only way out is a reboot. The more you look at BIOS
> calls and the complications, the worse it gets. I think that's one reason
> the Linux and *BSD folks are yanking ACPI up INTO the OS.
>
> I can see your point, but I think we'll be able to get what we want with
> the in-kernel ACPI support. It is pretty neat.
>
> ron
Hi
How about splitting the kernel in an HW dependent (immanent) part whih can and is
diretly bundeld and linked to linuxbios,
and a common systempart which needed to compiled only once.
After having finished the hw things the lowlevel kernel would try to find and load the
system kenel part which does all the rest needed by system and user.
Would be a bulk of work to get the kenel split into two parts without needing the
module overhead
cu
Christoph
--
THESIS: God is alive
PROOVE: Who else would have scheduled the mankind and world first
recommendation of resurch????
CONCLUSION: Scientist do what he wants, willing or not:)
