https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111875
--- Comment #7 from Andrew Pinski <pinskia at gcc dot gnu.org> --- (In reply to Richard Biener from comment #3) > git blame might tell - It was done for PR 56977. __builtin_constant_p folding in fab. > there was the alternative to > perform this propagation in fold_builtins but it's difficult to be > "complete" there. The alternative would be to turn that into a proper > simple constant propagation pass. > > Not sure if all worth for -Og just because of sanopt though. When I am finished moving everything away from fab into gimple-fold/forwprop (and other places), I will be replacing the fab/copyprop at -Og with a forwprop after pass_object_sizes. Since forwprop has its own simple copy-prop part to it and the difference between fab and forwprop is that fab only touches builtins while forwprop does all statements but the cost shouldn't be too much and forwprop is already done before inlining anyways so it is at -Og already.
