linux-kernel - Re: [RESEND PATCH] sched/fair: Skip sched_balance

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <e09fb1c2-cf74-47ff-af6c-671fe23c2e90@linux.ibm.com>
Date: Mon, 13 Oct 2025 22:11:05 +0530
From: Shrikanth Hegde <sshegde@...ux.ibm.com>
To: "Chen, Yu C" <yu.c.chen@...el.com>, Peter Zijlstra
 <peterz@...radead.org>,
        Tim Chen <tim.c.chen@...ux.intel.com>
Cc: Ingo Molnar <mingo@...nel.org>, Doug Nelson <doug.nelson@...el.com>,
        Mohini Narkhede <mohini.narkhede@...el.com>,
        linux-kernel@...r.kernel.org,
        Vincent Guittot <vincent.guittot@...aro.org>,
        K Prateek Nayak <kprateek.nayak@....com>
Subject: Re: [RESEND PATCH] sched/fair: Skip sched_balance_running cmpxchg
 when balance is not due



On 10/13/25 10:02 PM, Chen, Yu C wrote:
> On 10/13/2025 10:26 PM, Peter Zijlstra wrote:
>> On Thu, Oct 02, 2025 at 04:00:12PM -0700, Tim Chen wrote:
>>
>>> During load balancing, balancing at the LLC level and above must be
>>> serialized.
>>
>> I would argue the wording here, there is no *must*, they *are*. Per the
>> current rules SD_NUMA and up get SD_SERIALIZE.
>>
>> This is a *very* old thing, done by Christoph Lameter back when he was
>> at SGI. I'm not sure this default is still valid or not. Someone would
>> have to investigate. An alternative would be moving it into
>> node_reclaim_distance or somesuch.
>>
> 
> Do you mean the following:
> 
> diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
> index 444bdfdab731..436c899d8da2 100644
> --- a/kernel/sched/topology.c
> +++ b/kernel/sched/topology.c
> @@ -1697,11 +1697,16 @@ sd_init(struct sched_domain_topology_level *tl,
>                  sd->cache_nice_tries = 2;
> 
>                  sd->flags &= ~SD_PREFER_SIBLING;
> -               sd->flags |= SD_SERIALIZE;
>                  if (sched_domains_numa_distance[tl->numa_level] > 
> node_reclaim_distance) {
>                          sd->flags &= ~(SD_BALANCE_EXEC |
>                                         SD_BALANCE_FORK |
>                                         SD_WAKE_AFFINE);
> +                       /*
> +                        * Nodes that are far away need to be serialized to
> +                        * reduce the overhead of long-distance task 
> migration
> +                        * caused by load balancing.
> +                        */
> +                       sd->flags |= SD_SERIALIZE;
>                  }
> 
> We can launch some tests to see if removing SD_SERIALIZE would
> bring any impact.
> 
>>> On a 2-socket Granite Rapids system with sub-NUMA clustering enabled
>>> and running OLTP workloads, 7.6% of CPU cycles were spent on cmpxchg
>>> operations for `sched_balance_running`. In most cases, the attempt
>>> aborts immediately after acquisition because the load balance time is
>>> not yet due.
>>
>> So I'm not sure I understand the situation, @continue_balancing should
>> limit this concurrency to however many groups are on this domain -- your
>> granite thing with SNC on would have something like 6 groups?
>>
> 
> My understanding is that, continue_balancing is set to false after
> atomic_cmpxhg(sched_balance_running), so if sched_balance_domains()
> is invoked by many CPUs in parallel, the atomic operation still compete?
> 

 From what i could remember,

This mostly always happens at SMT after which continue_balancing = 0.
Since multiple CPUs end up calling it (specially on busy system)
it causes a lot cacheline bouncing. and ends up showing in perf profile.