That's wrong. All firmware I worked with used assembly for early initialization (startup.s) in those cases where the only alternative would be to rely on UB. I've worked with firmware that only worked with -O0 and failed with -O2, but these were always bad implementations, not due to some inherent requirement to rely on UB (most common example: failing to recognize that using volatile does not guarantee a memory barrier).
> I've worked with firmware that only worked with -O0 and failed with -O2
But as we can see in this vuln, there's functionality that works in the normal case but where important checks are elided, and compilers might decide not to do anything with a piece of undefined behaviour in one version but not in the next. Whether a piece of code works in -O2 doesn't prove it doesn't contain undefined behaviour.
