https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122473
--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> --- The reason why we warn is because we "over" vectorize the code and then expose that part for the warning. Now there needs to have a better way of handling these cases in general but there is no design for that. we could add an trap for when the out of bounds happen and not warn do you want that? Or do you want a warning where you might have missed that the argument input could be causing an out of bound case even if it is not very consistent? Also with the call case, well that needs full static anlysis power which this is out side of the scope of this warning; that is NOT out side of the scope of -fanalyzer though.
