Re:Spectre mitigations murder userspace performance
by ghoward on Sunday June 20, 2021 @10:38AM
(#27570451)
Fair points.
Regarding "formally verified," I believe the "formal" refers to the formal in formal methods, but I agree that it can be weaselly.
In the case of seL4, however, I don't think it is because they are incredibly precise about defining their assumptions. They have verified the machine code for various architectures, subject to those assumptions.
CompCert has not actually been formally verified - only the middle-end has. The C preprocessor and assembler aren't, the frontend formerly wasn't (but I think is now), and bugs have been found in those parts by fuzzing.
Plus being a faithful implementation of the C spec certainly leaves room for security issues…
(Why do we say "formally verified"? It sounds like weasel words. What is formal about it?)