basically you're misunderstanding or misrepresenting the kernel-side gap feature. like i said, this CVE will go away, it's undefendable.