+ //
+ // Note [upd-black-hole]
+ // One slight hiccup is that the THUNK_SELECTOR machinery can
+ // overwrite the updatee with an IND. In parallel GC, this
+ // could even be happening concurrently, so we can't check for
+ // the IND. Fortunately if we assume that blackholing is
+ // happening (either lazy or eager), then we can be sure that
+ // the updatee is never a THUNK_SELECTOR and we're ok.
+ // NB. this is a new invariant: blackholing is not optional.