There is either a legal need for WP to provide an opt-out if/when it relicense under a later (major revision) version of GFDL, or there isn't.
If there is, then we can stop talking about hypothetical, and talk about whether it'll be worth migrating at all, and if so how the opt out will work in practise.
If there isn't, then I would argue we don't provide an opt out under the very same reason we don't allow editors to retract their content submitted to WP now.
One can argue the extent of the efforts that would be required to rewrite articles to fulfil an opt out request, and it surely would depends on the number of such requests and whether the editor in concern would be willing to help. I would argue any reasonable estimates on the number of requests, and hence the number of articles affected would run into at least tens of thousands, probably more, as it would surely include all the number of editors we have had that tried to retract their submissions. Even if we are only talking about tenths of a percent of all articles on WP, we're still talking about a very disruptive number here.
Also, the idea that most if not all of those asking for the opt out will help in fulfilling those requests is, what's the phase, highly optimistic to say the least. Providing an opt out would be setting a very bad precedent to the future times when we have editors arguing for their contribution to be removed. The very simple argument that you let people removed their contribution in the past mean you should be allowing them to remove their contribution since then as well. This would imply disruption to potentially infinite number of articles for ever more.
KTC