Giammarco Zacheo wrote:
> Thank you. There's no hurry actually, just reporting the bug.
>
> If there's something I can do to help...
>
I have no efika board at hand, but just booted a plain lite5200 IceCube
successfully, with all tracing options enabled. This is a DENX-originated kernel
(tagged
Philippe Gerum wrote:
> Giammarco Zacheo wrote:
>> Thank you. There's no hurry actually, just reporting the bug.
>>
>> If there's something I can do to help...
>>
>
> I have no efika board at hand, but just booted a plain lite5200 IceCube
> successfully, with all tracing options enabled. This is a
Hello,
Antonio Del Cinque wrote:
> Hi,
> sorry, I needed time to implement your suggestions, but finally I got
> some results. I checked for cache related issues, and found that latency
> is highly related to cache misses.
>
> I found the way to disable the instruction cache (the e300 powerpc cor
Giammarco Zacheo wrote:
> It boots either with I-pipe support enabled and disabled.
>
> I can also boot with latency tracing enabled, if 'Instrument function
> entries' is not enabled.
>
Ok, so back to square #1. It's indeed a problem with mcount() referring to
unmapped memory at early init time.
We now found the reason for the really huge values and the complete hangup:
I am using a core 2 processor but didn't set the processor type within the
kernel configuration correctly.
Now I get good results again.
Regards
juergen
Posted by juergen . lanner on September 03, 2008 - 11:45: