http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58721
--- Comment #14 from Jan Hubicka <hubicka at ucw dot cz> --- > Actually, I would argue that the middle-end should be smart enough to give a > branch that is guaranteed to never return a negligible probability > (independent > of the builtin_expect). It can only be mis-predicted once. For predicting branches, we have gimple predict_stmt. If we need to annotate values with higher probability, I will implement the extension into bulitin_expect to handle them. (i.e. adding internal only second argument specifying the predictor) Sounds sane?
