[CostModel] remove cost-kind predicate for funnel shift costs

Completing the series of FIXME removals for special-case intrinsics:
50dfa19c
f2c25c70
c963bde0
01ea93d8

This one looks quite different than the others. The size/blended
cost is still potentially very far off from the throughput cost,
but this is hopefully not worse on the whole. It looks like the
underlying costs for the expanded shift/logic have their own
cost-kind limitations. Also, we are not asking the target if
it has a legal funnel shift op, so we just assume that the
intrinsic gets expanded.