Gerwin Klein

Overall it sounds like a good thing to reduce permissions. Why do you say it would be a breaking change? Because memory layout changes for the user if the kernel...

Right, so it's a potentially breaking change for boot loaders, i.e. before the kernel, not for user level. @nspin would that affect your Rust loader for instance?

I'm very happy with all new repos using `main` (which is what we are doing), but I am concerned with continuity on renaming the master branch in existing repos --...

Yup, can confirm no verification impact.

This really is weird. I had never looked at those times closely before. I also find it interesting that simulation for clang builds have consistently taken longer that the ones...

One thing to make progress on would be to confirm it is indeed that commit that makes the difference, e.g. by manually turning those options off and running the simulation...

This looks fine form my side, but it is a breaking API change and will need a short RFC first.

Yeah, ok, it's borderline, you're right that it's not really a new feature. I'm Ok to do this without RFC unless some from the @seL4/tsc objects. @yyshen do you have...

> One thing when considering the other invoke functions is that some of them actually can be preempted and return an exception, so those ones can't be changed to void....

> For `vcpu_fpu_to_host()`, only the complicated path is tested, as that is the default case when the host has FPU enabled. All in all I'm confident it's working for x86...