OK so impromptu, a large body of work landed in the partitioning crate. Primarily this surrounds partition table reloads (BLKPG),
and partition planning/strategies.
That's great, but we now need to link the data-centric APIs up to something useful, so that we can mutate partition tables and Do Things.
For now we're using gpt crate (MIT) which is a significant improvement over the 3.x series.