I submitted a theoretical paper about new algorithms to a computer science conference. Now I found out that the most important theorem of my paper was actually proved 20 years ago (in a workshop paper with less than 15 citations). The other theorems/algorithms of my paper are kind of easy corollaries of this result, so this makes the theoretical contribution of my paper lower. However my paper connects these results to a lot of literature of past 20 years to which the old paper has not been connected before, so it still has good motivation and a lot of original contribution.
What should I do if the reviewers do not find this old paper and accept my submission? Could an accept turn into decline if I tell them about this old paper? Would it be ok to just add a footnote where this old paper is cited?