INADA Naoki: "LOAD_METHOD support was based on tp_fastcall. (...) Other ideas to support LOAD_METHOD for C function are: * Add _PyMethodDescr_FastCallKeywords and call it in call_function."
It seems like a very small subset of tp_fastcall. If it provides a significant speedup, I really like the idea :-) I would be happy to see a "subset of tp_fastcall" merged. At least my whole tp_fastcall experiment wouldn't be worthless :-)
What would be the main change which would provide a speedup? Avoid the creation of a temporary tuple to pass positional arguments?