RMD should throw error if user unmount resctrl

[Shihta]

Because RMD should not cache/store any resctrl information, RMD should throw error if user unmount resctrl

Here are 2 cases where RMD should have responded to an error but did not

A. GET /workloads

1. Run RMD in a fresh environment, there is no workload/COS in RMD/resctrl
2. Create a workload via POST /workloads
3. umount resctrl
4. Query workloads via GET /workloads,
   and then you'll see that the response includes the workload you just created

B. GET /cache/l3

1. Run RMD in a fresh environment, there is no workload/COS in RMD/resctrl
2. Query cache information via GET /cache/l3
3. umount resctrl
4. Query cache information via GET /cache/l3,
   and then you'll get a response includes cache information without any Internal Server Error.