[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <8bf57eed-0e07-4102-900e-e63c1110815c@redhat.com>
Date: Mon, 3 Nov 2025 16:22:18 -0500
From: Waiman Long <llong@...hat.com>
To: Chen Ridong <chenridong@...weicloud.com>, Tejun Heo <tj@...nel.org>,
 Johannes Weiner <hannes@...xchg.org>, Michal Koutný
 <mkoutny@...e.com>
Cc: cgroups@...r.kernel.org, linux-kernel@...r.kernel.org,
 Chen Ridong <chenridong@...wei.com>, Gabriele Monaco <gmonaco@...hat.com>,
 Frederic Weisbecker <frederic@...nel.org>
Subject: Re: [cgroup/for-6.19 PATCH 2/3] cgroup/cpuset: Fail if isolated and
 nohz_full don't leave any housekeeping
On 11/2/25 10:53 PM, Chen Ridong wrote:
>
>
>> +
>>   	/*
>>   	 * Change the parent's effective_cpus & effective_xcpus (top cpuset
>>   	 * only).
>> @@ -2994,7 +3055,11 @@ static int update_prstate(struct cpuset *cs, int new_prs)
>>   		 * A change in load balance state only, no change in cpumasks.
>>   		 * Need to update isolated_cpus.
>>   		 */
>> -		isolcpus_updated = true;
>> +		if ((new_prs == PRS_ISOLATED) &&
>> +		    !isolated_cpus_can_update(cs->effective_xcpus, NULL))
>> +			err = PERR_HKEEPING;
>> +		else
>> +			isolcpus_updated = true;
>>   	} else {
>>   		/*
>>   		 * Switching back to member is always allowed even if it
> This is an issue I am trying to fix, the prstate_housekeeping_conflict check is necessary.
>
> https://lore.kernel.org/cgroups/20251025064844.495525-2-chenridong@huaweicloud.com/
>
You are right. We should add prstate_housekeeping_conflict() check here.
BTW, I found some issues when I looked further at the patch. So I am 
going to rewrite part of it and will send out a v2 after some testing.
Cheers,
Longman
Powered by blists - more mailing lists