Re:Plausibly deniable encryption (2020)
by josephcsible on Friday February 18, 2022 @03:29PM
(#30390658)
The whole concept of nuke keys is snake oil unless they're implemented with something like a TPM or HSM, since otherwise the attacker can just restore the pre-nuke image if you set it off.
[ Parent
| on HN ]
Re:Plausibly deniable encryption (2020)
by cookiengineer on Saturday February 19, 2022 @01:27AM
(#30394852)
While I agree with you with the snakeoil argument, I also have to say:
TPM chips are useless because there's already tools available to listen on the SPI bus for encryption keys. [1] [2]
It's only a matter of time until other key formats (other than Bitlocker) are supported.
I think the only way to secure the drive is being able to write your own drive controller firmware. But even then HDD plates could be just "swapped into" another HDD.
Re:Plausibly deniable encryption (2020)
by josephcsible on Saturday February 19, 2022 @12:14PM
(#30398636)
That attack isn't relevant in this case, because it only lets you steal the keys if the TPM chooses to release them to the CPU, but if you don't know the right PIN, or use a nuke PIN, then it won't do so.