[OLPC-devel] Secure BIOS on the OLPC

[Mark J. Foster]

Ivan Krstić wrote:
> When we do a power-off (i.e. init 0), the next start is a cold boot,
> correct?
>   
For the sake of discussion, let's say yes (though it isn't actually the 
same at all).  Even then, the vast majority of the time, folks will 
simply be doing suspend/resume.
>> On second thought, is this really better than making kids hold down the
>> spacebar for 4-5 seconds?
>>     
>
> Yes. It still provides secure upgrades without requiring user
> interaction.
Can we rely on waiting for folks to manually power-down their machines, 
when the power button is actually just a suspend/resume button?  To make 
this scheme work, we really need a hardware condition which software 
cannot replicate.  That's why I mention removing the battery, or 
*possibly* holding down the power button for four seconds (this would 
have to be investigated).

By the way, I initially missed Carl-Daniel's note regarding this topic, 
so consider my thoughts just an elaboration regarding what it would take 
to re-enable writes after we've disabled them.