> For example, if we allow processes to remain below their minimum run-queue
> provided they don't ever use up an entire timeslice, then a malevolent
> program could then intentionally try to always use 95% of its timeslice,
> but not the whole thing so as to avoid ever being sent back to its minimum
> run-queue. For a process like this, starvation could be the best thing
> that ever happened to it... a one-way ticket to run-queue 0 :-)
No it isn't. Starvation only occurs when there are other
processes competing for CPU all the time, and these
processes have a higher priority.
I think I have another solution for the problem. We
demote the problem to it's original minimum level when
it uses the CPU slice completely, and we demote it two
levels down when it doesn't completely finish it's CPU
quantum.
That way, a +20 niced vi will remain somewhat responsive,
but a +20 niced CPU intensive process won't get anything
more than it's share.
Rik.
+-----------------------------+------------------------------+
| For Linux mm-patches, go to | "I'm busy managing memory.." |
| my homepage (via LinuxHQ). | H.H.vanRiel@fys.ruu.nl |
| ...submissions welcome... | http://www.fys.ruu.nl/~riel/ |
+-----------------------------+------------------------------+
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu