The Old and Middle English g (the only other instance across Germanic) stands for [j] in this case, and Holtzmann's law doesn't apply to this root either. Here (upper right) this g is explained as secondary, developed from ǥ that emerged in its turn in the hiatus between o and o
Update. Hellquist regards a stage with w as secondary as well, but w was etymological there, so, if it was present in Old Swedish (is it attested?), it may have been original.