Security

5278 readers

1 users here now

Confidentiality Integrity Availability

founded 5 years ago

MODERATORS

AMD Zen 1 through Zen 4 CPUs use an insecure hash function in the signature validation for microcode updates; researchers released a proof of concept update which makes the RDRAND instruction return 4 (github.com)

submitted 1 month ago* (last edited 1 month ago) by [email protected] to c/[email protected]

8 comments fedilink hide all child comments

for readers missing the significance of the number 4 in the proof of concept: to demonstrate this vulnerability the researchers created a microcode update which replaces the "hardware" random number generator behind the RDRAND instruction with an implementation of xkcd#221 😭

you are viewing a single comment's thread
view the rest of the comments

[–] [email protected] 3 points 1 month ago (1 children)

In practical terms, can someone explain what this means? Ring 0 from outside a VM.

That would mean that if I were to have an image of an OS, as long as I have local admin while loading this on a VM, I would be able to run code as root?

[–] [email protected] 1 points 2 weeks ago (1 children)

Pretty sure it means the equivalent of root on baremetal (ie ring 0 in a vm is still safe).

Pretty sure on top of that anyone with that access can essentially rewrite your CPU's brain, allowing ring 0 access even after (for example) selling the CPU.

[–] [email protected] 1 points 2 weeks ago (1 children)

Didn't think anyone was gonna get back to me on this, thanks Tiddy.

If you had ring 0 on the bare metal, isn't it safe to assume that anything that bare metal has, including installed vms is also compromised or easily compromisable?

I'm not too knowledgeable on security yet but this interested me

[–] [email protected] 1 points 2 weeks ago

I thought the same, looks like AMD is trying to introduce something to limit that access (ie allow potentially compromised hosts run trusted VMs).

Probably to make VPS' more attractive to security focused divisions.