PyModule_GetState() gives you *per-module* state, not per-interpreter state.

Module objects are shared across subinterpreters, unless you use multi-phase initialization.

> PyModule_GetState() requires having the module object that corresponds
> to the given interpreter state.  I'm not sure how a C extension module
> is supposed to get its own module object corresponding to the current
> interpreter state, without getting it from the caller in some way.

This is the problem described in PEP 573: you don't always have access to your own module object. That keeps some more complex modules from switching to multi-phase init.

Unless this issue can wait for when PEP 580, PEP 573, and possibly some fallout of unknown unknowns are solved, let's add PyInterpreterState_GetDict for now.