https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123092
--- Comment #13 from Vladimir Makarov <vmakarov at gcc dot gnu.org> --- > Vlad -- how is this supposed to work? It seems wrong to be slamming in the > equivalence without either validating the result to verify its valid or > sending it back through curr_insn_transform to reload the invalid stuff. > The equivalence change should result in processing the insn on its constraints again. And in this case it did not happen. I've reproduced the bug. I think the fix will be ready tomorrow.
