Matthew Garrett Explains How to Increase Security at Boot Time

Who thinks there's excess anxiety in all these boot risks ? At least for John Doe's machine, 98% of any user base.

I'm finding myself agreeing with you nmset.

1. If you don't have control of a Linux system then you can't force a reboot.

2. If you are physically located at the system then you can force a reboot.

3. But we all know that if your system is physically compromised then you've already lost control of your system. We keep our systems physically secure.

So the paranoia over boot exploits seems to me also to be excessive. What am I missing?

Maybe this is a spillover from the windows mentality where rebooting a system is a normal part of the setup.exe installation procedure. It's the "When in doubt, reboot" mentality. Over here in the Linux world we have two good reasons to reboot: to switch to a new kernel, and to test our bootup scripts.

Best i can tell, its a government/military certification thing.

And this in turn bleeds into the wider _sec world.

And there paranoia is not a illness, but a work requirement.

Best i can tell, their basic policy is not to sleep, not to hibernate, but to shut a system down fully when not in use. Laptops in particular. And then it need to boot up into a verified good state before you enter your unlock codes for the encrypted and masked file systems.

This what leads to the twin pathologies of fretting over rapid boot (that the cloud/container-heads are also drooling over, aka "uptime by machinegun") and security in the face of "state level actors".

>Best i can tell, its a government/military certification thing.

That would rightly account for the remaining 2% or less. But such work requirements imply also that anything about it should not go public. Any way, I just wanted to spell out a "bored and fed up" state.