(In reply to Michal Suchanek from comment #5) > Right, so the -Wno-unknown-warning-option is misleading because gcc only > complains about it when there is other stuff to complain about. > > Why is the other error reported only on ppc64 then? No idea, probably "bad" luck due to different inlining/optimization decisions. > The code is equally broken on every arch. So its "good" luck then ;)