[okl4-developer] More NSLU2 debugging

[Geoffrey Lee]

On Tue, Jun 24, 2008 at 10:53:38AM +1000, Joshua Root wrote:
>> while (start_bootmem_phys == 0);
>> printf("shouldn't get here\n");
>>
>> Right now I don't have any good ideas other than trying to narrow down 
>> the problem a bit more and via code inspection.
>
> Further experimentation has revealed that the values I see in the  
> variables (correct vs 0) change with the number of print function calls  
> in the file, not with where I check them. Looks like something somewhere  
> is clobbering the stack.
>
> - Josh


Josh - 

Yes, I agree that looks like stack corruption and possibly register
corruption as well.  Are you using your own version of a print function,
or have you modified the platform serial driver to do early prints?
If you have some assembly in between, perhaps you have modified the
stack and/or registers in some unintended manner?

	-gl