https://gcc.gnu.org/bugzilla/show_bug.cgi?id=38093
--- Comment #4 from Eric Gallager <egallager at gcc dot gnu.org> --- So, if I understand the patch, isn't this just papering over the issue by fixing the specific instance of the bfin target causing this genrecog crash? That is, couldn't some other port also cause genrecog to crash similarly by making the same mistake? If that's the case, I think it'd make sense to patch genrecog to print a sensible error message instead of just crashing...
