http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48506
--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> 2011-04-07 21:53:33 UTC --- (In reply to comment #2) > I'd do the checkin, if you let me know if you prefer the one case removed, or > volatile added. Hmm, this testcase works for me in 4.6.x and on the trunk.