The problem is that even if there is no practical exploit, you still have to patch. At our web company where there are no shared boxes we still spent untold re

The problem is that even if there is no practical exploit, you still have to patch. At our web company where there are no shared boxes we still spent untold resources upgrading kernels, rebooting servers, and then lost capacity due to performance degradations. Infrastructure is normally patched and rebooted on a regular schedule, but when a bug like this hits and requires an off-cycle upgrade it's a big cost.