[OLPC-devel] Secure BIOS on the OLPC

[Carl-Daniel Hailfinger]

Ivan Krstić wrote:
> Ronald G Minnich wrote:
>> in the case that you changed your mind, for whatever reason,  but the
>> reboot happened before you had a chance to unwind the decision to flash.
> 
> In what circumstances do you foresee an OLPC user changing his mind
> about flashing the BIOS? This is the same question I asked Daniel. It's
> not rhetorical.
> 
>> If the OLPC folks are happy with the method you have outlined, then go
>> for it. I am expressing my opinion, based on several years of bios work
>> on thousands of machines, that it makes me uncomfortable.
> 
> I have full respect for your expertise in this area, which is why I'm
> trying to get to the bottom of what specifically makes you uncomfortable
> about it. To repeat the question from my previous e-mail, how is it
> worse than asking the user first (modulo the asking vs. not asking itself)?

User (or sysadmin) prepares BIOS update, writes it to specified file.
Battery is low so the reboot is scheduled to happen when a stable power
source is connected. Now the machine crashes and reboots automatically.
Battery runs out during flashing because there was no way to avoid
flashing in that situation. You now have a very nice brick.


Regards,
Carl-Daniel
-- 
http://www.hailfinger.org/