So, the TL;DR seems to be that ~all major CPU vendors fail at side channels in some way, but Intel is failing way more than the others by allowing userland<->kernel infoleaks. Correct?
Yeah, but there is "leak within a single process' address space" (couldn't care less about that, can be mitigated with better browser sandboxing / hardened JITs) vs. "leak across processes' address space" / "leak across userland/kernel boundary" which are very different beasts.
-
-
the project 0 blog post has PoCs leaking kernel mem from user mode and host memory from guest without using the intel-specific issue
-
Without using eBPF-JIT? Did I miss one of the PoCs?
- 5 more replies
New conversation -
-
-
There are enough side channels and possible applications of them to create a large matrix, and I think not enough research has generally been done to answer about any really definitively (esp. in a generic way).
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Don’t forget “leak between VMs” and “leak between VM and hypervisor”
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.