Currently, _reportUpstream_inlock(unique_lock* lock) does not guarantee the state of lock upon return. This makes it harder to reason about code in which we subsequently invoke functions (such as updateTerm_incallback()) that expect to lock the mutex:

replication_coordinator_impl.cpp:443

    _reportUpstream_inlock(&lk);
    _externalState->setGlobalTimestamp(lastOpTime.getTimestamp());
    // Step down is impossible, so we don't need to wait for the returned event.
    _updateTerm_incallback(term);