[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Z4m1XqHl4whTG1pl@slm.duckdns.org>
Date: Thu, 16 Jan 2025 15:41:50 -1000
From: Tejun Heo <tj@...nel.org>
To: Changwoo Min <changwoo@...lia.com>
Cc: void@...ifault.com, arighi@...dia.com, kernel-dev@...lia.com,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH 5/7] sched_ext: Add an event, SCX_EVENT_RQ_BYPASSING_OPS
On Fri, Jan 17, 2025 at 12:15:41AM +0900, Changwoo Min wrote:
> Add a core event, SCX_EVENT_RQ_BYPASSING_OPS, which represents how many
> operations are bypassed once the bypass mode is set.
>
> Signed-off-by: Changwoo Min <changwoo@...lia.com>
> ---
> kernel/sched/ext.c | 23 +++++++++++++++++++++--
> 1 file changed, 21 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
> index 094e19f5fb78..44b44d963a0c 100644
> --- a/kernel/sched/ext.c
> +++ b/kernel/sched/ext.c
> @@ -1485,6 +1485,11 @@ struct scx_event_stat {
> * is dispatched to a local DSQ when exiting.
> */
> u64 ENQ_LOCAL_EXITING;
> +
> + /*
> + * When the bypassing mode is set, the number of bypassed operations.
> + */
> + u64 RQ_BYPASSING_OPS;
> };
I'm not sure this is a particularly good way to account for bypass mode.
Maybe account the total duration of bypass modes, the number of times it was
activated and the number of times tasks were dispatched in bypass mode?
Thanks.
--
tejun
Powered by blists - more mailing lists