A clustered index will sort row data in a table according to the ordering of the fields in the index. If so, why would you (or would you not) want to create a clustered index on the child table in a parent/child relationship?

Lets assume the classical order-with-orderlines example where you never fetch the orderlines without first going through the associated order. I would think that having a clustered index on both the orderid and the orderlineid column (as clustered indexes must be unique) forces all orderlines for a single order to be physically located adjacent to each other, which would make reads very performant. Is this correct? And if so, are there downsides to this approach?