Stefan Bambach <[EMAIL PROTECTED]> writes:

[snip]

> I think that the INT functions from the normal BIOS code is the real
> problem to port this kind a project for other OS. Older Windows are
> based on DOS and that is based on BIOS INT functions. So here you have
> a problem. But why not thinking of BSD, Solaris (Intel), BeOS or some
> sort of Unix/Posix variant. They do not need the BIOS much. So the
> BIOS startup time can be minimized.

I don't know what the slow part is but I don't think it is the INT
calls.  But since we don't support then concentrating on OS that don't
heavily depend on the is probably a good idea.

> I think it would be nice for every computer user to speed up the
> complete process from power on to windows login, bash prompt, ... .
> 
> Measuring time needed for complete process, it clear that the OS only
> needs about half the time to load and the other half is the BIOS.
> About double speed is much time you save.

The one feature we haven't yet gotten a good implementation of is
waiting for hard driver spinup (or telling the OS to wait for us).
As far as I know that is the only thing that could put a real crimp in
our boot time.

> 
> What is the problem booting other OS ?

No ports...  There are no fundamental reasons LinuxBIOS can't support
another OS.

> The lack of BIOS INT functions ?
Yes.

> Do actual OS like W2000, XP, ... need them as DOS did ?
Not nearly as much.  But their boot process still relies on them.

Eric

Reply via email to