ctwentz
Re:UC Berkeley to build open-source secure enclave using RISC-V
on Friday December 14, 2018 @03:04PM (#18684146)
Some version at least will be yes. RE: RoT + RISV-V, cache isolation appears missing here, so sidechannel attacks using cache timing and other methods still work
Re:UC Berkeley to build open-source secure enclave using RISC-V
on Friday December 14, 2018 @12:36PM (#18682674)
Isn't this highly similar to MIT's Sanctum from 2015, also open source and ignoring the hardware changes needed to make meaningful guarantees about sidechannel attack robustness? Useful if you want to avoid hardware mods, but the enclave model is not new.
https://eprint.iacr.org/2015/564.pdf and open source code via MIT license here: https://github.com/pwnall/sanctum
Re:Apple T2 Security Chip: Security Overview [pdf]
on Tuesday October 30, 2018 @06:07PM (#18341695)
The T2 is utilizing a physically unclonable function (PUF) and TRNG to create a public/private device key pair on die. At manufacturing, the T2 exposes the device public key to Apple, Apple signs with a group key posted to their CA. This is why they say they can revoke privileges via the CA. It's unclear to me whether the T2 regenerates this key pair each time it is requested, or whether it is encrypted and stored in memory. In the event the latter is the case, the encryption is being performed inside the secure enclave. "Secure enclave" used here is almost certainly distinct from Intel SGX, ARM TrustZone and the like. The page tables are protected. Row Hammer, Spectre, Meltdown and Foreshadow do not apply to something like T2, as the OS is considered trusted. The fundamental challenge that e.g. Intel's SGX has over this type of architecture is that a dedicated security co-processor doesn't need to maintain speculative execution behaviors necessitated by performance requirements, which expose numerous sidechannel attacks, and likely has minimal need to assume untrusted code operating in the T2 OS.