Setting the policy

[cs412017]

While setting the policy for which mode to be in MLF / RR, from the
scenarios, we can infer that while moving from RR to MLF, we have to
only move the idle thread to queue number 3. When moving from MLF to RR,
we have to concatenate all the queues into queue 0.

However, should we do the same when the user sets the policy to be MLF
when the policy is already MLF? That is, should we still move all the
threads to queue 0 and only leave the idle thread in queue 3?

It seems to me that we shouldn't do anything in this case...that is,
just don't do anything.

Is this assumption right?

[Iulian Neamtiu]

That's right, you don't have to move threads around if you get a
Set_Scheduling_Policy(P,Q) if the current policy is P. However, you might
have to adjust the quantum, if the current quantum is not Q.

Iulian

Post by cs412017
While setting the policy for which mode to be in MLF / RR, from the
scenarios, we can infer that while moving from RR to MLF, we have to
only move the idle thread to queue number 3. When moving from MLF to RR,
we have to concatenate all the queues into queue 0.
However, should we do the same when the user sets the policy to be MLF
when the policy is already MLF? That is, should we still move all the
threads to queue 0 and only leave the idle thread in queue 3?
It seems to me that we shouldn't do anything in this case...that is,
just don't do anything.
Is this assumption right?