http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47967
--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-03-04 13:36:50 UTC --- Created attachment 23543 --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23543 gcc46-pr47967.patch Untested fix. Or is there some better place to check for this earlier? Inlining certainly handles this right and doesn't consider inlining such functions, but IPA-CP didn't care.
