Robin Rosenberg [off-list ref] writes:
One could consider keeping the contributions from ^1 a special case and not
mention the parent, making it look like any revert commit. I guess most merge
reverts are like this in practice.
I think that makes sense. There are cases where the mainline maintainer
punts a merge and pass the baton to a subsystem maintainer, saying "Your
tree has many conflicts with my tip, and I'd rather ask you to resolve it"
(and after such a merge, the mainline maintainer will fast forward to the
result), in which case the merge will be in the reverse direction, but
that should be rare. Reverting such a merge later from the mainline's
point of view would involve "revert -m 2".
So if your patch is tightened a bit to record extra information only in
such a case, I think that would be an acceptable approach to the issue.