Trusted platform module security defeated in 30 minutes, no soldering required
arstechnica.com
arstechnica.com
TL;DR - The laptop owner didn't use a PIN on the TPM or a secondary credential. The full disk encryption key was read off a SPI bus from the TPM.
So? It seems to me this is a big problem for the default out of the box user experience.
Edit: Ok after searching the article I can't tell how/where they sealed the key. Bitlocker tells the TPM to measure some things and if that measurement matches was in PCR12 (or whatever is defined) then reveal a secret. That secret is the bitlocker key.
(Personally I'd love to see the encryption and key storage module being a physically separable device that I could carry with me. Something like an HSM with I/O bandwidth sufficient to perform the bulk encryption on the device itself.)
Clever, and certainly feels like a design flaw that should be addressed.
But also not something most people are going to be able to pull off. If you have physical access to the computer, it's always going to be game over. And, if you don't, social engineering always works if you try hard enough.
It would be really nice (If anybody from MSFT is reading this!) if they encrypted that key w/ a public key. At this point at least, they just store it there in plaintext.
Solution: use fTPM (integrated into the CPU)
Not if it's off and you don't leave the key around.
Why did no one get suspicious when they noticed that they were not even being asked for a PIN?
Disk encryption key was in cleartext on the SPI bus