mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Chen, Yu C" <yu.c.chen@intel.com>
To: Aaron Lu <ziqianlu@bytedance.com>
Cc: Tim Chen <tim.c.chen@linux.intel.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@redhat.com>,
	K Prateek Nayak <kprateek.nayak@amd.com>,
	"Gautham R . Shenoy" <gautham.shenoy@amd.com>,
	Vincent Guittot <vincent.guittot@linaro.org>,
	Juri Lelli <juri.lelli@redhat.com>,
	Dietmar Eggemann <dietmar.eggemann@arm.com>,
	"Steven Rostedt" <rostedt@goodmis.org>,
	Ben Segall <bsegall@google.com>, Mel Gorman <mgorman@suse.de>,
	Valentin Schneider <vschneid@redhat.com>,
	"Madadi Vineeth Reddy" <vineethr@linux.ibm.com>,
	Hillf Danton <hdanton@sina.com>,
	"Shrikanth Hegde" <sshegde@linux.ibm.com>,
	Jianyong Wu <jianyong.wu@outlook.com>,
	"Yangyu Chen" <cyy@cyyself.name>,
	Tingyin Duan <tingyin.duan@gmail.com>,
	Vern Hao <vernhao@tencent.com>, Len Brown <len.brown@intel.com>,
	Aubrey Li <aubrey.li@intel.com>, Zhao Liu <zhao1.liu@intel.com>,
	Chen Yu <yu.chen.surf@gmail.com>,
	Adam Li <adamli@os.amperecomputing.com>,
	Tim Chen <tim.c.chen@intel.com>, <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 10/19] sched/fair: Prioritize tasks preferring destination LLC during balancing
Date: Wed, 29 Oct 2025 21:19:54 +0800	[thread overview]
Message-ID: <253f00cc-2bea-43b6-aca3-96105014c4be@intel.com> (raw)
In-Reply-To: <CANCG0GdR-0Yh16nuAWq477RyWXhzbo_ppDsH9hQrbkxRqw1+-g@mail.gmail.com>

On 10/29/2025 5:51 PM, Aaron Lu wrote:
> On Mon, Oct 27, 2025 at 10:00:52AM +0800, Chen, Yu C wrote:
>> Hi Aaron,
>>
>> On 10/24/2025 5:32 PM, Aaron Lu wrote:
>>> Hi Tim,
>>>
>>> On Sat, Oct 11, 2025 at 11:24:47AM -0700, Tim Chen wrote:
>>>> @@ -10849,11 +10849,45 @@ static void record_sg_llc_stats(struct lb_env *env,
>>>>    	if (unlikely(READ_ONCE(sd_share->capacity) != sgs->group_capacity))
>>>>    		WRITE_ONCE(sd_share->capacity, sgs->group_capacity);
>>>>    }
>>>> +
>>>> +/*
>>>> + * Do LLC balance on sched group that contains LLC, and have tasks preferring
>>>> + * to run on LLC in idle dst_cpu.
>>>> + */
>>>> +static inline bool llc_balance(struct lb_env *env, struct sg_lb_stats *sgs,
>>>> +			       struct sched_group *group)
>>>> +{
>>>> +	struct sched_domain *child = env->sd->child;
>>>> +	int llc;
>>>> +
>>>> +	if (!sched_cache_enabled())
>>>> +		return false;
>>>> +
>>>> +	if (env->sd->flags & SD_SHARE_LLC)
>>>> +		return false;
>>>> +
>>>> +	/* only care about task migration among LLCs */
>>>> +	if (child && !(child->flags & SD_SHARE_LLC))
>>>> +		return false;
>>>> +
>>>> +	llc = llc_idx(env->dst_cpu);
>>>> +	if (sgs->nr_pref_llc[llc] > 0 &&
>>>> +	    can_migrate_llc(env->src_cpu, env->dst_cpu, 0, true) == mig_llc)
>>>
>>> llc_balance() is called from update_sg_lb_stats() and at that time,
>>> env->src_cpu is not determined yet so should not be used here?
>>>
>>
>> You are right, I think we should check the candidate group's first
>> CPU rather than the env->src_cpu. Will fix it in the next version.
> 
> Looks like can_migrate_llc() doesn't care an exact cpu but any cpu in the
> same LLC should do, so either the candidate group's first cpu or any
> other cpus in that group should make no difference.
> 

Yes, actually the cache aware is based on LLC rather than CPU. It was
a historic reason that the original proposal was based on CPU in the
wakeup path.

> It might be more intuitive to prototype can_migrate_llc() with sd_shared
> as param than using cpu, just a thought.

I understand your concern,  will think about this.

thanks,
Chenyu


  reply	other threads:[~2025-10-29 13:20 UTC|newest]

Thread overview: 116+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-10-11 18:24 [PATCH 00/19] Cache Aware Scheduling Tim Chen
2025-10-11 18:24 ` [PATCH 01/19] sched/fair: Add infrastructure for cache-aware load balancing Tim Chen
2025-10-14 19:12   ` Madadi Vineeth Reddy
2025-10-15  4:54     ` Chen, Yu C
2025-10-15 19:32       ` Tim Chen
2025-10-16  3:11         ` Chen, Yu C
2025-10-15 11:54     ` Peter Zijlstra
2025-10-15 16:07       ` Chen, Yu C
2025-10-23  7:26   ` kernel test robot
2025-10-27  4:47   ` K Prateek Nayak
2025-10-27 13:35     ` Chen, Yu C
2025-10-11 18:24 ` [PATCH 02/19] sched/fair: Record per-LLC utilization to guide cache-aware scheduling decisions Tim Chen
2025-10-15 10:15   ` Peter Zijlstra
2025-10-15 16:27     ` Chen, Yu C
2025-10-27  5:01   ` K Prateek Nayak
2025-10-27 14:07     ` Chen, Yu C
2025-10-28  2:50       ` K Prateek Nayak
2025-10-11 18:24 ` [PATCH 03/19] sched/fair: Introduce helper functions to enforce LLC migration policy Tim Chen
2025-10-11 18:24 ` [PATCH 04/19] sched/fair: Introduce a static key to enable cache aware only for multi LLCs Tim Chen
2025-10-15 11:04   ` Peter Zijlstra
2025-10-15 16:25     ` Chen, Yu C
2025-10-15 16:36       ` Shrikanth Hegde
2025-10-15 17:01         ` Chen, Yu C
2025-10-16  7:42           ` Peter Zijlstra
2025-10-17  2:08             ` Chen, Yu C
2025-10-16  7:40       ` Peter Zijlstra
2025-10-27  5:42   ` K Prateek Nayak
2025-10-27 12:56     ` Chen, Yu C
2025-10-27 23:36       ` Tim Chen
2025-10-29 12:36         ` Chen, Yu C
2025-10-28  2:46       ` K Prateek Nayak
2025-10-11 18:24 ` [PATCH 05/19] sched/fair: Add LLC index mapping for CPUs Tim Chen
2025-10-15 11:08   ` Peter Zijlstra
2025-10-15 11:58   ` Peter Zijlstra
2025-10-15 20:12     ` Tim Chen
2025-10-11 18:24 ` [PATCH 06/19] sched/fair: Assign preferred LLC ID to processes Tim Chen
2025-10-14  5:16   ` Chen, Yu C
2025-10-15 11:15     ` Peter Zijlstra
2025-10-16  3:13       ` Chen, Yu C
2025-10-17  4:50       ` Chen, Yu C
2025-10-20  9:41         ` Vern Hao
2025-10-11 18:24 ` [PATCH 07/19] sched/fair: Track LLC-preferred tasks per runqueue Tim Chen
2025-10-15 12:05   ` Peter Zijlstra
2025-10-15 20:03     ` Tim Chen
2025-10-16  7:44       ` Peter Zijlstra
2025-10-16 20:06         ` Tim Chen
2025-10-27  6:04   ` K Prateek Nayak
2025-10-28 15:15     ` Chen, Yu C
2025-10-28 15:46       ` Tim Chen
2025-10-29  4:32         ` K Prateek Nayak
2025-10-29 12:48           ` Chen, Yu C
2025-10-29  4:00       ` K Prateek Nayak
2025-10-28 17:06     ` Tim Chen
2025-10-11 18:24 ` [PATCH 08/19] sched/fair: Introduce per runqueue task LLC preference counter Tim Chen
2025-10-15 12:21   ` Peter Zijlstra
2025-10-15 20:41     ` Tim Chen
2025-10-16  7:49       ` Peter Zijlstra
2025-10-21  8:28       ` Madadi Vineeth Reddy
2025-10-23  6:07         ` Chen, Yu C
2025-10-11 18:24 ` [PATCH 09/19] sched/fair: Count tasks prefering each LLC in a sched group Tim Chen
2025-10-15 12:22   ` Peter Zijlstra
2025-10-15 20:42     ` Tim Chen
2025-10-15 12:25   ` Peter Zijlstra
2025-10-15 20:43     ` Tim Chen
2025-10-27  8:33   ` K Prateek Nayak
2025-10-27 23:19     ` Tim Chen
2025-10-11 18:24 ` [PATCH 10/19] sched/fair: Prioritize tasks preferring destination LLC during balancing Tim Chen
2025-10-15  7:23   ` kernel test robot
2025-10-15 15:08   ` Peter Zijlstra
2025-10-15 21:28     ` Tim Chen
2025-10-15 15:10   ` Peter Zijlstra
2025-10-15 16:03     ` Chen, Yu C
2025-10-24  9:32   ` Aaron Lu
2025-10-27  2:00     ` Chen, Yu C
2025-10-29  9:51       ` Aaron Lu
2025-10-29 13:19         ` Chen, Yu C [this message]
2025-10-27  6:29   ` K Prateek Nayak
2025-10-28 12:11     ` Chen, Yu C
2025-10-11 18:24 ` [PATCH 11/19] sched/fair: Identify busiest sched_group for LLC-aware load balancing Tim Chen
2025-10-15 15:24   ` Peter Zijlstra
2025-10-15 21:18     ` Tim Chen
2025-10-11 18:24 ` [PATCH 12/19] sched/fair: Add migrate_llc_task migration type for cache-aware balancing Tim Chen
2025-10-27  9:04   ` K Prateek Nayak
2025-10-27 22:59     ` Tim Chen
2025-10-11 18:24 ` [PATCH 13/19] sched/fair: Handle moving single tasks to/from their preferred LLC Tim Chen
2025-10-11 18:24 ` [PATCH 14/19] sched/fair: Consider LLC preference when selecting tasks for load balancing Tim Chen
2025-10-11 18:24 ` [PATCH 15/19] sched/fair: Respect LLC preference in task migration and detach Tim Chen
2025-10-28  6:02   ` K Prateek Nayak
2025-10-28 11:58     ` Chen, Yu C
2025-10-28 15:30       ` Tim Chen
2025-10-29  4:15         ` K Prateek Nayak
2025-10-29  3:54       ` K Prateek Nayak
2025-10-29 14:23         ` Chen, Yu C
2025-10-29 21:09         ` Tim Chen
2025-10-30  4:19           ` K Prateek Nayak
2025-10-30 20:07             ` Tim Chen
2025-10-31  3:32               ` K Prateek Nayak
2025-10-31 15:17                 ` Chen, Yu C
2025-11-03 21:41                   ` Tim Chen
2025-11-03 22:07                 ` Tim Chen
2025-10-11 18:24 ` [PATCH 16/19] sched/fair: Exclude processes with many threads from cache-aware scheduling Tim Chen
2025-10-23  7:22   ` kernel test robot
2025-10-11 18:24 ` [PATCH 17/19] sched/fair: Disable cache aware scheduling for processes with high thread counts Tim Chen
2025-10-22 17:21   ` Madadi Vineeth Reddy
2025-10-23  6:55     ` Chen, Yu C
2025-10-11 18:24 ` [PATCH 18/19] sched/fair: Avoid cache-aware scheduling for memory-heavy processes Tim Chen
2025-10-15  6:57   ` kernel test robot
2025-10-16  4:44     ` Chen, Yu C
2025-10-11 18:24 ` [PATCH 19/19] sched/fair: Add user control to adjust the tolerance of cache-aware scheduling Tim Chen
2025-10-29  8:07   ` Aaron Lu
2025-10-29 12:54     ` Chen, Yu C
2025-10-14 12:13 ` [PATCH 00/19] Cache Aware Scheduling Madadi Vineeth Reddy
2025-10-14 21:48   ` Tim Chen
2025-10-15  5:38     ` Chen, Yu C
2025-10-15 18:26       ` Madadi Vineeth Reddy
2025-10-16  4:57         ` Chen, Yu C

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=253f00cc-2bea-43b6-aca3-96105014c4be@intel.com \
    --to=yu.c.chen@intel.com \
    --cc=adamli@os.amperecomputing.com \
    --cc=aubrey.li@intel.com \
    --cc=bsegall@google.com \
    --cc=cyy@cyyself.name \
    --cc=dietmar.eggemann@arm.com \
    --cc=gautham.shenoy@amd.com \
    --cc=hdanton@sina.com \
    --cc=jianyong.wu@outlook.com \
    --cc=juri.lelli@redhat.com \
    --cc=kprateek.nayak@amd.com \
    --cc=len.brown@intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mgorman@suse.de \
    --cc=mingo@redhat.com \
    --cc=peterz@infradead.org \
    --cc=rostedt@goodmis.org \
    --cc=sshegde@linux.ibm.com \
    --cc=tim.c.chen@intel.com \
    --cc=tim.c.chen@linux.intel.com \
    --cc=tingyin.duan@gmail.com \
    --cc=vernhao@tencent.com \
    --cc=vincent.guittot@linaro.org \
    --cc=vineethr@linux.ibm.com \
    --cc=vschneid@redhat.com \
    --cc=yu.chen.surf@gmail.com \
    --cc=zhao1.liu@intel.com \
    --cc=ziqianlu@bytedance.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®