https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=265974
--- Comment #11 from Mark Johnston <ma...@freebsd.org> --- I do think the deferred advance might still be incorrect on platforms where FreeBSD's atomic_thread_fence_rel() does not provide a store/load barrier, such as POWER. -- You are receiving this mail because: You are the assignee for the bug.