On 05/04/13 08:35, Artyom Tarasenko wrote:
Actually there is another option - does the previous allocation before this one that returns 0xffc67000 as the virtual address? In that case it could indicate that va=0x0 doesn't have a special meaning after all.
Is it possible for you to post the complete log output from an unpatched OpenBIOS for us to look at so we can see the complete pattern of allocations?
The with the unpatched one it dies pretty early, right after claiming va=0x0. See the the attachment.
Got it. Hmmm definitely a special behaviour when va == NULL seems to be the more likely explanation for the moment. Let me know if anything changes with my latest patch.
Otherwise I think you'd have to disassemble the boot loader a bit to see where the va = NULL is coming from. If it's a constant 0x0 then that would definitely confirm that the value has a special behaviour.
ATB,
Mark.