I'm working on PPC PREP hardware. These systems are not like Macs, they
use a different boot procedure. A properly marked partition (dos style
part table) will be read right into memory and executed. AFAIK, the BIOS
does this directly. I don't think the MBR is looked at.
Now booting on the mac is a different issue. ;-)
I'd start with just PREP system support, but I would code with an eye
towards Mac support in the future. Perhaps Benjamin Herrenschmidt
<[EMAIL PROTECTED]> the author of BootX would get involved if I asked him.
BootX is the Mac OS extension that most mac distros use as a boot menu.
What would you think of typedefing the ints etc to like u16, u32 etc?
and then a header could insure the native system has the right bits.
If there is interest, I might start a proof of concept port.
OKUJI Yoshinori wrote:
>
> From: Tim Riker <[EMAIL PROTECTED]>
> Subject: Platform support?
> Date: Mon, 13 Mar 2000 17:36:58 -0700
>
> > How much work do you all think would be involved in taking
> > the existing ppc boot assembly code and patching grub to work there?
>
> I don't know, since I have little knowledge about any ppc
> platform. What platform are you working on? Mac?
>
> Anyway, I think it would be hard work because I've written quite
> x86-dependennt code even for the part in C. For example, I assume:
> sizeof(int) = 32, sizeof(long) = 32, sizeof(void *) = 32 so there is
> no actual difference among them, and the byte order is little
> endian. Probably the assumed memory layout would trouble you as
> well. Therefore, my current opinion is that GRUB *is* very
> unportable. This might be changed in the future when we will
> completely rewrite it (possibly after next public release).
>
> > PS: did you know that OpenLinux 2.4 uses grub as a loader? Have you seen
> > the graphical interface is uses for the selection screen?
>
> They are crazy! =)
>
> Thanks,
> Okuji
--
Tim Riker - http://rikers.org/ - short SIGs! <g>
All I need to know I could have learned in Kindergarten
... if I'd just been paying attention.
