I was thinking more along the lines of "re-initialization" .. particularly WRT memory
configs ... where the contents of memory get trashed on initialisation .. or are you
proposing to take something like boot parameters to sqelsh any attempts for hardware
re-configs by subsequent kernels .. ?
----- Original Message -----
>From: <[EMAIL PROTECTED]>, <Eric W. Biederman>
>To: "Ian" <[EMAIL PROTECTED]>
>Subject: Re: question on superio
>Date: Mon, 26 Feb 2001 17:13:50 -0700
>
> "Ian" <[EMAIL PROTECTED]> writes:
>
> > > 2) If you can get reliable kernel driver that always sets up hardware it
> > > doesn't matter who sets it up.
> > What happens when kernels are booted from kernels? Or when kernels are
> > software booted from wares such as VMWare? (I haven't ever run it, so I don't
> > know how it handles hardware manipulation -- nor do I want to see linuxbios
> > external to the kernel ..).
>
> The point here is that if you put the hardware into a proper state the
> kernel just works no matter if the hardware has been setup before.
>
> Kernels booted from kernels might get a benefit if the first kernel
> knew how to initialize the hardware and the second kernel didn't, it
> depends on how the reset the hardware.
>
> As for vmware it might mean it has to emulate a few extra hardware
> registers.
>
> But the point is if you unconditially set hardware to a known state in
> your driver you have more reliable code period. This works for
> linuxBIOS but it also works for other BIOS's as well. The framebuffer
> drivers are a good example of this.
>
>
> Eric
>
--
http://HumanHeuristic.com/
"Bringing people together in a world full of computers"
