https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126445
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> --- I'll note that we need to sit down and fix this on the costing side, with costmodel testcases. There's also still related patches pending which is why I didn't yet start to investigate any of these cases. Can we possibly link the same-rev-bisect bugs via see-also?
