Messages in this thread | | | Date | Fri, 3 Feb 2006 18:36:52 -0800 | From | "Siddha, Suresh B" <> | Subject | Re: [PATCH] load_balance: "busiest CPU" -> "busier CPUs" |
| |
On Fri, Feb 03, 2006 at 04:38:07PM -0800, hawkes@sgi.com wrote: > In these circumstances, an all-pinned "busiest CPU" will effectively > disable load_balance balancing.
Solving all the load balancing issues that occur under all-pinned case is tricy... For example, even with this patch, at a particular sched domain, load balance might still be disabled for the cpus which belong to the same sched group as the all-pinned "busiest CPU"
> @@ -243,6 +243,8 @@ struct runqueue { > int active_balance; > int push_cpu; > > + int cpuid; /* of this runqueue */ > + > task_t *migration_thread; > struct list_head migration_queue;
A simple change to find_busiest_queue() can avoid that addition to the runqueue struct.
thanks, suresh - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |