http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55079
--- Comment #6 from Jan Hubicka <hubicka at gcc dot gnu.org> 2012-11-02 18:44:37 UTC --- Hmm, it seems to be due to off-by-one bug in my patch Index: tree-ssa-loop-ivcanon.c =================================================================== --- tree-ssa-loop-ivcanon.c (revision 193098) +++ tree-ssa-loop-ivcanon.c (working copy) @@ -405,11 +405,11 @@ remove_exits_and_undefined_stmts (struct into unreachable (or trap when debugging experience is supposed to be good). */ if (!elt->is_exit - && elt->bound.ult (double_int::from_uhwi (npeeled))) + && elt->bound.ule (double_int::from_uhwi (npeeled))) { gimple_stmt_iterator gsi = gsi_for_stmt (elt->stmt); gimple stmt = gimple_build_call (builtin_decl_implicit (BUILT_IN_UNREACHABLE), 0); gimple_set_location (stmt, gimple_location (elt->stmt)); gsi_insert_before (&gsi, stmt, GSI_NEW_STMT); I however introduced it because w/o it we get bootstrap miscompare. Looks like I will need to debug it after all :(