Skip to content

Comment on SeL4 security proofs now complete on AArch64parent

Comments

I don't think you can ignore memory access. The whole "hyperthreading considered harmful" was because of shared cache between contexts. That's why I think a proof is in reach without cache, much harder with one...

Without cache, without virtual memory, without throttling when the more hardware is switched on or the CPU gets hot, without memory that’s unreliable under load (https://en.wikipedia.org/wiki/Row_hammer), etc.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.