https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81655
Jakub Jelinek <jakub at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |ASSIGNED Last reconfirmed| |2017-08-02 Assignee|unassigned at gcc dot gnu.org |jakub at gcc dot gnu.org Ever confirmed|0 |1 --- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Created attachment 41890 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41890&action=edit gcc8-pr81655.patch Untested fix for the trunk, the issue is that we have in GIMPLE_COND bool_var != 1 and record that as low == high == 1 with negated in_p, while the code expects low == high == 0 without negated in_p, which we would get e.g. for bool_var == 0.