https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116380
Richard Biener <rguenth at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Version|unknown |15.0 Status|NEW |ASSIGNED Assignee|unassigned at gcc dot gnu.org |rguenth at gcc dot gnu.org --- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> --- There's <bb 6> [local count: 201326592]: # e_25 = PHI <e_13(19), e_32(12)> # c_lsm.3_6 = PHI <(19), c_lsm.3_26(12)> # c_lsm_flag.4_21 = PHI <1(19), c_lsm_flag.4_27(12)> if (c_lsm_flag.4_21 != 0) goto <bb 7>; [66.67%] else goto <bb 8>; [33.33%] where the PHI value is missing on the path we pattern-matched a memset. It was a default-def and should have been left alone. Probably failure of the manual SSA updating code. I'll have a look.