It looks like the code to expand the swap in case of a memory size increase is already present in the existing code, and what is being enhanced is the inverse - reducing swap to follow a memory size decrease.
Rejecting from the queue because the changelog implies that we are making a change that is partially already present. The user impact needs adjusting also please, since it describes a problem that does not currently exist in the series in question.
This also raises the question of what happens if a user deliberately increases their swap. Will this reduce it again contrary to their expectations, or is there protection against that?
It looks like the code to expand the swap in case of a memory size increase is already present in the existing code, and what is being enhanced is the inverse - reducing swap to follow a memory size decrease.
Rejecting from the queue because the changelog implies that we are making a change that is partially already present. The user impact needs adjusting also please, since it describes a problem that does not currently exist in the series in question.
This also raises the question of what happens if a user deliberately increases their swap. Will this reduce it again contrary to their expectations, or is there protection against that?