Numerous vulnerabilities in AMD processors

Depending on whether whatever AMD's equivalent of Boot Guard is enabled, write access to the BIOS chip shouldn't be exploitable for anything other than a secure boot bypass and control over CPL0 and up. MASTERKEY (if the vulnerability is for real) gives SMM privilege. The degree to which this is a problem is admittedly rather dubious.

> 2) RYZENFALL: again, loading unauthorised code on the Secure Processor as admin. Threat level: No shit, Sherlock!

I disagree. The whole point of the PSP is that it should *not* be tamperable with as admin. This allows whatever TPM-like features it emulates to be compromised, SEV to be compromised, etc. OTOH, SEV is thoroughly insecure be design anyway, at least in current revisions.

I personally have no idea why MS and other users consider an emulated TPM to be a TPM at all for purposes of MS/Windows logo requirements, etc.

Previously, and paradoxically, you had to buy an *Intel* if you wanted an x86 that's so pitifully insecure that there were trivial automated white-hat tools to remove its known backdoors (me-cleaner etc.). The AMD PSP was considered airtight enough to be an actual threat to security (no user serviceable parts, but still IoT-on-a-chip garbage) but now the cat's been let out of the bag, there'll likely be enough eyeballs on the hardware to fix the problem.