Re: [PATCH] sched_ext: Take NUMA node into account when allocating per-CPU cpumasks

[Andrea Righi <arighi@nvidia.com>]

On Mon, Feb 10, 2025 at 04:52:25PM +0800, lirongqing wrote:
> From: Li RongQing <lirongqing@baidu.com>
> 
> per-CPU cpumasks are dominantly accessed from their own local CPUs,
> so allocate them node-local to improve performance.

Makes sense to me, did you run some tests/benchmarks on any large NUMA
system with this?

Thanks,
-Andrea

> 
> Signed-off-by: Li RongQing <lirongqing@baidu.com>
> ---
>  kernel/sched/ext.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
> index 8857c07..3fe5a2e 100644
> --- a/kernel/sched/ext.c
> +++ b/kernel/sched/ext.c
> @@ -6325,15 +6325,16 @@ void __init init_sched_ext_class(void)
>  
>  	for_each_possible_cpu(cpu) {
>  		struct rq *rq = cpu_rq(cpu);
> +		int  n = cpu_to_node(cpu);
>  
>  		init_dsq(&rq->scx.local_dsq, SCX_DSQ_LOCAL);
>  		INIT_LIST_HEAD(&rq->scx.runnable_list);
>  		INIT_LIST_HEAD(&rq->scx.ddsp_deferred_locals);
>  
> -		BUG_ON(!zalloc_cpumask_var(&rq->scx.cpus_to_kick, GFP_KERNEL));
> -		BUG_ON(!zalloc_cpumask_var(&rq->scx.cpus_to_kick_if_idle, GFP_KERNEL));
> -		BUG_ON(!zalloc_cpumask_var(&rq->scx.cpus_to_preempt, GFP_KERNEL));
> -		BUG_ON(!zalloc_cpumask_var(&rq->scx.cpus_to_wait, GFP_KERNEL));
> +		BUG_ON(!zalloc_cpumask_var_node(&rq->scx.cpus_to_kick, GFP_KERNEL, n));
> +		BUG_ON(!zalloc_cpumask_var_node(&rq->scx.cpus_to_kick_if_idle, GFP_KERNEL, n));
> +		BUG_ON(!zalloc_cpumask_var_node(&rq->scx.cpus_to_preempt, GFP_KERNEL, n));
> +		BUG_ON(!zalloc_cpumask_var_node(&rq->scx.cpus_to_wait, GFP_KERNEL, n));
>  		init_irq_work(&rq->scx.deferred_irq_work, deferred_irq_workfn);
>  		init_irq_work(&rq->scx.kick_cpus_irq_work, kick_cpus_irq_workfn);
>  
> -- 
> 2.9.4
>