As long as we're speculating on what sort of things might be handled with a hash-table mechanism that is covered by the patent we're concerned about, I should add the case of UUIDs. UIIDs (GUIDs to some of us) are quite heavily used these days and hashed-table lookups for them are not uncommon.
Whether and how that might figure in Linux use of the patented method is not anything I have a clue about. And without knowing what the actual cases are, it is difficult to speculate further about what would be a suitable work-around.
I presume that the Linux kernel developers are doing what they can to figure that out.