[Shared Everything] struct.wait on unshared should trap

[kripken]

No description provided.

[kripken]

@tlively looks like the spec test errors on this PR, as it has a wait on a non-shared struct. Is the spec test wrong...?

[stevenfontanella]

The spec test was written by me. I think the spec isn't clear on what should happen with a non-shared waitqueue, this is all it says:

Since the spec test cases never actually need to be woken up by another thread, I could see it making sense for a non-shared waitqueue but I'm not sure if it's useful to allow it at all. e.g. one of the cases is that the control word doesn't match so the thread doesn't block.

Anyway the test isn't from the proposal so feel free to change it if needed.

[tlively]

Yeah, we should update the README to be more precise (probably as part of the ongoing redesign of waitqueues), but this is to match the analogous behavior of the existing wait instructions.

[kripken]

Sounds good @stevenfontanella @tlively , I updated the spec test.

[tlively]

It would be good to at least add a TODO for adding a test for the trap-on-unshared behavior.

[kripken]

I added a full spec test and removed the exec test.