> Users who are concerned about the flaw have no other recourse than to run the entire OS as a properly configured virtual machine. Looks like OS can mitigate

> Users who are concerned about the flaw have no other recourse than to run the entire OS as a properly configured virtual machine. Looks like OS can mitigate the problem though? All it needs to do is to check if the bits have changed just before context switch, and possibly in other interrupt handlers for a good measure. If they did, the current process is killed and security message is logged. This is not going to be 100% mitigation, but it should be enough to block most communications methods, especially if there are checks in hard-to-predict interrupts.