Follow

"Fault injection attacks against AMD's Secure Encrypted Virtualization" is this morning's highly recommended breakfast reading arxiv.org/pdf/2108.04575.pdf

@jakob @th Looks lovely. I've heard of the Remarkable devices.

@th So one question I have if about the voltage glitching. They say it doesn't require physical access to the machine, but when I read that chapter (3.3), it seems to describe only methods that do require such physical access. Do I understand correctly that such voltage regulators can be accessed remotely?

@mplammers their technique requires hardware access (look at all those wires and relays!), although other fault injection papers have demonstrated remote vulnerabilities through unprotected channels to the onboard voltage regulators. Plundervolt, for instance, causes faults in SGX enclaves and leaks key material plundervolt.com/

@th @mplammers It's kind of fun, it does need physical access - although only SPI+VR+PSU reset; but does need a few hours access - but then you walk away with what you need; it's kind of the opposite the other attack this week which doesn't need physical access but you suck a few bits at a time.

@th
That photo looks like a 3D render!

Your morning routine looks slick 😉

@th How do you get your PDF's on your RM2? Do you do that manual by connecting it to your PC, or do you use some kind of service?

@th
I guess this is kind of to be expected, at least for a few generations of the technology, but nonetheless I applaud AMD's efforts to build an actual high performance trusted execution environment.

Sign in to participate in the conversation
(void *) social site

(void*)