Re: [Gumstix-users] CPU Frequency scaling -- boot error

[Alexandre P. N. <al...@om...>]

>Thank you both, your input is greatly appreciated.  Ok, I understand now
>that switching the pxa255 CPU freq. is a tricky and unreliable thing to
>do.  But I don't understand how that explains the boot/load failure that
>I am experiencing.  It is not a matter of the system crashing because I
>attempted to have my user app. change the CPU frequency.  I can't even
>get past bootup. 
> 
>Can you please elaborate on that a bit more?
>
>Thank you!!
>
>Mark
>  
>

The conservative semantics is self-sustained (what I mean is it needs no 
intervention from the userspace), so it probably switches frequencies by 
itself during boot.  That could mess with several things and that may be 
the reason why the system can't boot.

But it could be something else completely unrelated. I remember to 
experience something similar when I recompiled the kernel after changing 
some parameter. But it was not the changed parameter that caused that, 
instead it was the simple recompiling fact, because some mtd driver had 
a linking bug (some object file needed to be bound before or after 
another). I think by that time saving the kernel's .config file to the 
default used by the buildroot and completely removing the linux tree, 
thus forcing a recompile from scratch seemed to solve my problem at this 
time.

I think that was already fixed on recent mtd trees, but I'm not sure 
what happened since then since I have not compiled kernels with 
nondefault configs since then.

Good luck figuring that out!

Alexandre