>On Mon, 11 Feb 2002 at 17:43:16, you wrote: >(ref: <046201c1b324$13798490$5c15ac3e@cobra>) > >>To paraphrase Phoebus, all you need to do is rewrite the (S)GC boot >>code :-) > >Indeed, and understand why they patched the ROMs. >OK John, nice of you to volunteer (8-)# > >BTW Stuart Honeyball hasn't a clue - it was all done by his side kick >(whose name I forget).
Mike? A couple of years ago I got excited by the possibility of porting uCLinux to the (S)GC. This is what I figured out before I got tired of wrestling with the uncommented disassembled code and undocumented memory-mapped I/O: 1) Just like the OS ROM, what you see when you look at the (S)GC ROM on a running system is not what you get if you read the ROM using an EPROM programmer 2) On start-up, the (S)GC ROM occupies the bottom 64K of the 680x0 memory map and, just like the OS ROM, has the reset SSP and PC values in locations 0 and 4. 3) The reset PC points to the (S)GC boot code which lives in the 48K - 64K region 4) The boot code copies the (S)GC ROM and the OS ROM into RAM, overwrites the start of the (S)GC ROM image with the standard QDOS ROM header and patches the OS ROM image to cope with >1MB RAM etc. 5) At some point, the processor does a soft reset and, as if by magic, the patched ROM images appear in their familiar places in the memory map! And no, I'm not volunteering for anything :-) (The last time I did that, I ended up on the QUANTA committee!) Regards, John -- [EMAIL PROTECTED]
