@k-sareen: Reminder to talk about flushing mutator buffers before GC (including destroyed mutators)

@k-sareen: Reminder to talk about flushing mutator buffers before GC (including destroyed mutators)

#1047 suggested that the binding should not need to flush mutator explicitly, i.e. the binding should call destroy_mutator but not flush_mutator. The contract is that the binding should let MMTk know when a mutator thread dies, and MMTk deals with whatever needs to be done with the mutator (e.g. flushing thread local states).

@qinsoon This is the PR I was talking about. It's a bit outdated now I think it we don't have the coordinator thread anymore, for example.