[Bug ipa/60911] [4.9/4.10 Regression] wrong code with -O2 -flto -fipa-pta

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60911

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
Seems like simd_clone_create gets along with this because it does

      if (!cgraph_function_with_gimple_body_p (old_node))
        return NULL;
      cgraph_get_body (old_node);
      new_node = cgraph_function_versioning (old_node, vNULL, NULL, NULL,
                                             false, NULL, NULL, "simdclone");

and cgraph_function_versioning applies IPA transforms?