BUG: slicing DatetimeIndex with unused and unordered levels gives incorrect results

[phofl]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

midx = MultiIndex(levels=[Index(["2019-12-31", "2019-12-30"], dtype="datetime64[ns]")], codes=[[0, 0]])
ser = Series([1, 2], index=midx)
ser.loc[:"2019-12-30"]

Issue Description

2019-12-31    1
2019-12-31    2
dtype: int64

This came out from the other issue, the index things it's monotonic, but this does not translate to it's levels.

cc @jbrockmendel thoughts on where to fix this? We could patch the indexing op, but not sure whether we could potentially run into other issues

Expected Behavior

empty series

Installed Versions

main

[jbrockmendel]

Looks like this goes through slice_locs -> get_slice_bound -> _partial_tup_index, but it isn't immediately obvious which of those holds the problem