https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64149
--- Comment #2 from Richard Earnshaw <rearnsha at gcc dot gnu.org> --- Sounds sensible to me. We switched to LRA quite late in gcc-4.9, so keeping a way to switch back in case of problems was pragmatic. But we've been running with the new code now for a year and not encountered any major issues that couldn't be fixed pretty quickly.
