Re: better late than never: amigappc progress

[Frank Wille <frank%phoenix.owl.de@localhost>]

Jukka Andberg wrote:

> Looks like the memory is there, but only first 8K at 0xfff00000 is
> writable.

Ok. So just the part of the exception area which is required by a 603e was
left writable.


> Is it necessary to have the boot code up there?

No. The area is initialized by the kernel later, so we can also put the boot
code at 0xfff01000, for example. I'll change that.


> Anyway after running your BPPCFix tool ("bppcfix install reboot") all
> of it becomes writable.

Which means the ROM-based ppc.library protects this area on every reboot.


>  So then I can run gobsdppc a bit further. It
> now prints Booting Single-User...Using 32256K FASTMEM at 0x76000000, 2M
> CHIPMEM then resets the display and nothing more happens.
> Which I guess is pretty much what can be expected currently.

Yes. The kernel is compiled for a base address of 0x8000000 and cannot
start.


> 32M in first slot and 4M in second slot:
> Memory Address: $66000020  Size: $0237FFE0 (36351K)     Type=$00
> CACHEABLE

Ok... at least it's always contiguous and never crosses a 256MB page border.


> And one more: 4M in first slot and another 4M in second slot:
> Memory Address: $47C00020  Size: $0077FFE0 (7679K)      Type=$00
> CACHEABLE

LOL
So every address between 0x40000000 and 0x80000000 is possible. But no
problem as long as the memory stays in a single 256MB page.


I will try to experiment with relocation techniques in the next days, and
you will know when there is a first test kernel for the A1200.

-- 
Frank Wille