Follow

Wow, some very interesting new research here: tpm.fail/

Researchers published attacks that can recover encryption keys from Trusted Computing Modules (TPMs) (which isn't supposed to be possible) by timing elliptic curve signature generation. "On Intel fTPM, our key recovery succeeds after about1,300 observations and in less than two minutes."

Nearly all Windows computers store BitLocker encryption keys in TPMs.

Β· Β· Web Β· 6 Β· 79 Β· 49

@micahflee I was under the impression this attack only affects one model of TPM chips

@maryjane under "Am I affected by these vulnerabilities?" the websites says, "There is a high chance that you are affected. This depends if any of your computing devices (laptop, tablet, desktop, etc.) use Intel fTPM or STMicroelectronics TPM chips."

@micahflee

I know Qubes uses TPM for AEM, some people close to me have enabled it. Personally, I'd prefer to disable TPM, TXT as far as possible.

Out of curiousity, would you care to elaborate why you have stopped using AEM?

@h3artbl33d it's not trivial to setup, and at some point I formatted my disk on that Qubes computer and re-installed, and decided it wasn't worth it. I'd protect against evil maid attacks by other means -- tamper evidence, stuff like the Haven app, etc.

@micahflee

Thank you for your reply, I understand that. Haven is a useful tool even more when combined with other means to protect against "evil maids". I learned from Trammel Hudson how well a bit of nail polish can hold up ;)

@h3artbl33d I went through a glitter nail polish phase myself. It's pretty cool, but the problem I had was comparing a nail-polished screw with a photo I took earlier. It's really hard to be 100% sure that nothing changed, especially when the lighting is completely different

@micahflee

No sole method is 100% safe. Thus, a combination is often safer. I use a variety with the systems housing sensitive data (nail polish, Haven, and more).

From a friend whom is more into nail polosh, I have understood that this one (flickr.com/photos/osr/15773860) is bad and could potentionally be circumvented with the right tools and some caution.

Might be fun to test this out :)

CW: computer security paranoia bait 

@micahflee It's almost as if cryptographers know what they're talking about when they warned us over and over and over again that this would happen...

Nah.

@micahflee I wonder can this be used to crack open any DRM schemes, before it's patched.. πŸ€”

@micahflee AFAIK all of the TPM chips communicate unencrypted with the CPU. Given two to three weeks, it is therefore possible to get whatever key the TPM is holding by sniffing it off the bus. See pulsesecurity.co.nz/articles/T for details.

So in the end it boils down to different attack scenarios, as usual. If you want to protect against the casual thief, then you're probably fine with BitLocker and TPM. But don't rely on it for anything more...

Sign in to participate in the conversation
Mastodon

Server run by the main developers of the project 🐘 It is not focused on any particular niche interest - everyone is welcome as long as you follow our code of conduct!