Sukadev Bhattiprolu <sukadev@xxxxxxxxxxxxxxxxxx> writes: > The obvious trade-off is with systems that don't use containers which > are porbably the majority at present. For them next_task() now becomes > more expensive (instead of a simply going to next item on list, they have > lookup in the pidmap, a lookup in pid hash table followed by mapping the > pid back to task). I think there was a discussion once on this and the > conclusion was things like "kill sig -1" are inherently expensive. Cost wise it would be worth measuring. I have a report that when that change was made to /proc readdir in /proc sped up. The problem that I see is that changing generic methods is not generally applicable. > Do you need these to be optimized for containers for some other reason ? A good question. Eric _______________________________________________ Containers mailing list Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/containers