https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119484
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> --- Created attachment 60894 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=60894&action=edit This is the reduced testcase from PR 119376 This is the represenative of the reduced testcase from PR 119376. The only difference between this one and the previous one is h is always inlined into p and that causes us to do: ``` <bb 4> [local count: 108662672]: goto <bb 6>; [100.00%] <bb 5> [local count: 108662672]: f.constprop.isra (); goto <bb 4>; [100.00%] <bb 6> [local count: 1073741824]: # _2 = PHI <0(4), 1(2)> return _2; ``` Which is how the IR looks for the one of the failures of PR 119376 .
