Re: cpusets not cpu hotplug aware

From: Anton Blanchard
Date: Wed Aug 23 2006 - 10:58:59 EST


> Well... let's suck it and see (please). If for some reason it proves
> inadequate and the default kernel behaviour is significantly wrong (it
> seems to be) then there's an argument for modifying (ie: adding complexity
> to) the kernel.

I think there is. We have a userspace visible change to the sched
affinity API when the cpusets option is enabled. Papering over it with a
udev callback doesnt sound like the right solution.

Im struggling to understand why we have this problem at all. If a task
has not been touched by cpuset calls it should be allowed to use any
cpu. I completely agree that once you have partitioned the task with
cpusets then it should never spill onto more recently hotplug added

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at