[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87a7d1d0-4b13-4f0f-afd6-58cb3a040ada@quicinc.com>
Date: Mon, 28 Jul 2025 17:44:36 +0800
From: Zhongqiu Han <quic_zhonhan@...cinc.com>
To: K Prateek Nayak <kprateek.nayak@....com>, <mingo@...hat.com>,
<peterz@...radead.org>, <juri.lelli@...hat.com>,
<vincent.guittot@...aro.org>, <dietmar.eggemann@....com>,
<rostedt@...dmis.org>, <bsegall@...gle.com>, <mgorman@...e.de>,
<vschneid@...hat.com>
CC: <linux-kernel@...r.kernel.org>, Zhongqiu Han <quic_zhonhan@...cinc.com>
Subject: Re: [PATCH] sched: Combine ENQUEUE_MIGRATED check in activate_task()
On 7/25/2025 3:14 PM, K Prateek Nayak wrote:
> Hello Zhongqiu,
>
> On 7/24/2025 5:42 PM, Zhongqiu Han wrote:
>> Combine the task_on_rq_migrating() check and the sched_mm_cid_migrate_to()
>> invocation into a single conditional block. This removes a redundant flag
>> check and slightly reduces the instruction count in a hot path.
>>
>> Signed-off-by: Zhongqiu Han <quic_zhonhan@...cinc.com>
>> ---
>> - This is a slight optimization and code cleanup, please let me know if you'd prefer to keep the current structure.
>> - Thanks
>>
>> kernel/sched/core.c | 4 ++--
>> 1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> index 2343f5691c54..d6063cf503ee 100644
>> --- a/kernel/sched/core.c
>> +++ b/kernel/sched/core.c
>> @@ -2118,10 +2118,10 @@ inline bool dequeue_task(struct rq *rq, struct task_struct *p, int flags)
>>
>> void activate_task(struct rq *rq, struct task_struct *p, int flags)
>> {
>> - if (task_on_rq_migrating(p))
>> + if (task_on_rq_migrating(p)) {
>> flags |= ENQUEUE_MIGRATED;
>> - if (flags & ENQUEUE_MIGRATED)
>
> I think "ENQUEUE_MIGRATED" can be individually set without
> task_on_rq_migrating() in the wakeup path via ttwu_do_activate().
>
> In that case "p->on_rq", goes from "0" to "TASK_ON_RQ_QUEUED" directly
> without going through an intermediate "TASK_ON_RQ_MIGRATING" but
> "ENQUEUE_MIGRATED" will be set if "p->task_cpu" was changed during
> wakeup.
>
Hi Prateek
You're right. Thanks for the nice review~
--
Thx and BRs,
Zhongqiu Han
Powered by blists - more mailing lists