https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120443
--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> --- So the reason for the original testcase in comment #0 started to fail on the trunk is either the inline or IPA-CP change which allowed to happen even into cold function calls (aka main). The one in comment #2 just renames main to f (marks it as noinline) and then calls f from main. This by passes the change that was done on the trunk to get a testcase to fail previously.
