Ronald G Minnich <[EMAIL PROTECTED]> writes:
> On 2 Jun 2001, Eric W. Biederman wrote:
>
> > I was messing around with the boot process, and managed
> > to slightly modify mkrom and start32.inc so I can
> > successfully use rombase 0xfffff000. Basically
> > it simply required using a short jump instead of a long
> > one and using %cs instead of %es in loading the gdt.
>
> are you going to commit this. It sounds good. We'll have to test on a lot
> of chipsets, and it should be optional, since we've found some that don't
> have the fffff000 turned on when the chipset turns on.
>
> Or did you fix that issue too :-)
There must be some amount of code length of code turned on
when the chipset turns on. Maybe it is naive to think that 64K
is mapped. However this only affects mkrom.
Currently I do a near jump if the rom is 64K or less long. This
might need to get modified. Everything else is fine. I just
use %cs instead of %es in start32.inc for lgdt and don't even play
with %es.
I play a lot when I'm developing code so I implement everything make
certain it is working (at least for me), and then submit it back into
the mainline. I've a ways to got but I'll probably start making
commits within the next work or two.
Eric
