From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.remote-ip=115.124.28.194 ARC-Seal: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; cv=none; t=1793784865; b=RHsYKKuBfJlyT7zx2rMUvincqrO7jScaW6jEDKN+pfcDdMQxPmlU4Ma4aGIfyJa7+BvE eHai9Lv/HE3aYRfAH4RHpO2Mi5E60iA7y6f02As2xhCljprneizI01BEM4yJO6wrkMjoE Q1p3x49N16vQIvzNvhG+eHm9JKf8yDqeTnIgHlTm5aIkOVrX1J3QejEnkp5Vuyp6yclS4 TH9YfOvS3cpgNa3YuV81hI2jIwvgQk5ZFnHb2wYc+xIP51HIodklpjgJ3Gt4n7BRhtYZ+ IJRVoIaWpVsH2z6vE4KohxXgpbd4thSipbtnL6iingKKj+ydFe11V+n3dW9ZUcgPJIA== ARC-Message-Signature: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; c=relaxed/relaxed; t=1793784865; h=DMARC-Filter:X-Alimail-AntiSpam:Received:Subject:To:Cc:References: From:Message-ID:Date:User-Agent:MIME-Version:In-Reply-To:Content-Type: Content-Transfer-Encoding; bh=/hbPdXBha6nBDFQ8LAmK466OmnWOluaenC9AL0XNqTs=; b=ti6YAn0BBMMDICMvmG0zMJIzIyOUTBsJfmdM/qkg9APTNbGG01nJf41zThAhG2e8bPYn ntXJ6iWOsaIKGAveIpJuBJKPmRRuVcmBESH6WGDM/E+hC9WpodWG7R2wSW4kvYwgPli/Q AVo2q8/oXutZ0UesRh9Mus7iTt/NSgDQPyshfeGFds5X5oJ9fZzPzzsFE7U1ngEHRsK9n horZmB9ozze6BGsO4515/NE8y8/pY2OIoo8WQexP75qwOmTd2zXAGW5prr1G4wjpmEbXR bRsn7klXh7uVpTva69J7BzPewzmZVcxgFhv+QGyovdJbRb0H06W5Za/VOQh4einJsvw== ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none header.from=open-hieco.net; spf=pass smtp.mailfrom=open-hieco.net; arc=none smtp.remote-ip=115.124.28.194 Received: from out28-194.mail.aliyun.com (out28-194.mail.aliyun.com [115.124.28.194]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (prime256v1) server-digest SHA256) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 22577606A331 for ; Wed, 4 Nov 2026 09:34:22 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.2 smtp.subspace.kernel.org 22577606A331 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=open-hieco.net Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=open-hieco.net X-Alimail-AntiSpam:AC=CONTINUE;BC=0.07436259|-1;CH=green;DM=|CONTINUE|false|;DS=CONTINUE|ham_regular_dialog|0.142377-0.0232716-0.834351;FP=11153015351543416059|0|0|0|0|-1|-1|-1;HT=maildocker-contentspam033023087102;MF=likunyu@open-hieco.net;NM=1;PH=DS;RN=11;RT=11;SR=0;TI=SMTPD_---.jwgrKdR_1793784852; Received: from 172.18.25.145(mailfrom:likunyu@open-hieco.net fp:SMTPD_---.jwgrKdR_1793784852 cluster:ay29) by smtp.aliyun-inc.com; Wed, 04 Nov 2026 17:34:13 +0800 Subject: Re: [PATCH] sched/fair: Distribute burst-woken tasks when rq depth reaches 3x average To: mingo@redhat.com, peterz@infradead.org, juri.lelli@redhat.com, vincent.guittot@linaro.org, dietmar.eggemann@arm.com, rostedt@goodmis.org, bsegall@google.com, mgorman@suse.de, vschneid@redhat.com, kprateek.nayak@amd.com Cc: linux-kernel@vger.kernel.org References: <20261104091842.266273-1-likunyu@open-hieco.net> From: Li kunyu Message-ID: Date: Wed, 4 Nov 2026 17:34:12 +0800 User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:78.0) Gecko/20100101 Thunderbird/78.11.0 X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Post: List-Help: List-Subscribe: List-Unsubscribe: List-Owner: Precedence: list MIME-Version: 1.0 In-Reply-To: <20261104091842.266273-1-likunyu@open-hieco.net> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit On 2026/11/4 17:18, Li kunyu wrote: > When a task burst-wakes a large number of tasks in a short period, tasks > can pile up on a single CPU regardless of whether the system has idle > CPUs or NO_HZ enabled. > > Track the average runqueue depth (scaled by SCHED_CAPACITY_SCALE to > preserve fractional precision) in root_domain during the periodic > load-balance pass. > > In select_task_rq_fair(), when the target runqueue depth reaches 3x the > system-wide average, fall back to cpumask_any_and_distribute(p->cpus_ptr, > cpu_active_mask) to rotate waking tasks across allowed CPUs and reset the > periodic load balance intervals so regular load balancing can promptly > smooth out the distribution. > > Suggested-by: Peter Zijlstra > Link: https://lore.kernel.org/all/20261021125419.GK2823843@noisy.programming.kicks-ass.net/ > Signed-off-by: Li kunyu > --- > kernel/sched/fair.c | 70 ++++++++++++++++++++++++++++++++++++++--- > kernel/sched/sched.h | 3 ++ > kernel/sched/topology.c | 2 ++ > 3 files changed, 71 insertions(+), 4 deletions(-) > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c > index 6e005f1e8dc9..e4d278ac46c4 100644 > --- a/kernel/sched/fair.c > +++ b/kernel/sched/fair.c > @@ -9903,6 +9903,38 @@ static int find_energy_efficient_cpu(struct task_struct *p, int prev_cpu) > return target; > } > > +/* > + * Check if a CPU's rq depth reaches 3x the system-wide average tracked > + * in root_domain (scaled by SCHED_CAPACITY_SCALE, floored at 1 task). > + */ > +static inline bool rq_burst_overloaded(int cpu) > +{ > + struct rq *rq = cpu_rq(cpu); > + unsigned long avg; > + > + if (unlikely(!rq->rd)) > + return false; > + > + avg = max_t(unsigned long, READ_ONCE(rq->rd->avg_nr_running), > + SCHED_CAPACITY_SCALE); > + > + return (READ_ONCE(rq->nr_running) << SCHED_CAPACITY_SHIFT) >= avg * 3; > +} > + > +static inline void reset_periodic_lb_interval(int cpu) > +{ > + struct sched_domain *sd; > + > + /* > + * Only the most basic values are updated here to avoid excessive > + * cost. > + */ > + for_each_domain(cpu, sd) { > + if (READ_ONCE(sd->balance_interval) > sd->min_interval) > + WRITE_ONCE(sd->balance_interval, sd->min_interval); > + } > +} > + > /* > * select_task_rq_fair: Select target runqueue for the waking task in domains > * that have the relevant SD flag set. In practice, this is SD_BALANCE_WAKE, > @@ -9972,11 +10004,19 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int wake_flags) > > /* Slow path */ > if (unlikely(sd)) > - return sched_balance_find_dst_cpu(sd, p, cpu, prev_cpu, sd_flag); > - > + new_cpu = sched_balance_find_dst_cpu(sd, p, cpu, prev_cpu, sd_flag); > /* Fast path */ > - if (wake_flags & WF_TTWU) > - return select_idle_sibling(p, prev_cpu, new_cpu); > + else if (wake_flags & WF_TTWU) > + new_cpu = select_idle_sibling(p, prev_cpu, new_cpu); > + > + if (unlikely(rq_burst_overloaded(new_cpu))) { > + int dist_cpu = cpumask_any_and_distribute(p->cpus_ptr, cpu_active_mask); > + > + if (likely(dist_cpu < nr_cpumask_bits)) { > + reset_periodic_lb_interval(new_cpu); > + return dist_cpu; > + } > + } > > return new_cpu; > } > @@ -11685,6 +11725,7 @@ struct sg_lb_stats { > struct sd_lb_stats { > struct sched_group *busiest; /* Busiest group in this sd */ > struct sched_group *local; /* Local group in this sd */ > + unsigned long total_nr_running; /* Sum of nr_running in this sd */ > unsigned long total_load; /* Total load of all groups in sd */ > unsigned long total_capacity; /* Total capacity of all groups in sd */ > unsigned long avg_load; /* Average load across all groups in sd */ > @@ -12964,6 +13005,7 @@ static inline void update_sd_lb_stats(struct lb_env *env, struct sd_lb_stats *sd > /* Now, start updating sd_lb_stats */ > sds->total_load += sgs->group_load; > sds->total_capacity += sgs->group_capacity; > + sds->total_nr_running += sgs->sum_nr_running; > > sum_util += sgs->group_util; > sg = sg->next; > @@ -12982,11 +13024,31 @@ static inline void update_sd_lb_stats(struct lb_env *env, struct sd_lb_stats *sd > env->fbq_type = fbq_classify_group(&sds->busiest_stat); > > if (!env->sd->parent) { > + unsigned int weight = env->sd->span_weight; > + > /* update overload indicator if we are at root domain */ > set_rd_overloaded(env->dst_rq->rd, sg_overloaded); > > /* Update over-utilization (tipping point, U >= 0) indicator */ > set_rd_overutilized(env->dst_rq->rd, sg_overutilized); > + > + /* > + * We update the average task count at the top-level (root) > + * domain, as this incurs the least overhead and comes > + * closest to the ideal scenario. The trade-off is that we > + * must accept some temporary data inaccuracy. > + */ > + if (weight) { > + unsigned long cur_avg; > + unsigned long old_avg; > + > + cur_avg = (sds->total_nr_running << SCHED_CAPACITY_SHIFT) / weight; > + old_avg = READ_ONCE(env->dst_rq->rd->avg_nr_running); > + if (old_avg) > + cur_avg = (old_avg * 7 + cur_avg) >> 3; > + > + WRITE_ONCE(env->dst_rq->rd->avg_nr_running, cur_avg); > + } > } else if (sg_overutilized) { > set_rd_overutilized(env->dst_rq->rd, sg_overutilized); > } > diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h > index e656c7059bf8..4147e043407c 100644 > --- a/kernel/sched/sched.h > +++ b/kernel/sched/sched.h > @@ -1077,6 +1077,9 @@ struct root_domain { > cpumask_var_t rto_mask; > struct cpupri cpupri; > > + /* Scaled (SCHED_CAPACITY_SCALE) average per-CPU rq depth */ > + unsigned long avg_nr_running; > + > /* > * NULL-terminated list of performance domains intersecting with the > * CPUs of the rd. Protected by RCU. > diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c > index 3dab0253976f..41c3b9e30d65 100644 > --- a/kernel/sched/topology.c > +++ b/kernel/sched/topology.c > @@ -558,6 +558,8 @@ static int init_rootdomain(struct root_domain *rd) > > if (cpupri_init(&rd->cpupri) != 0) > goto free_cpudl; > + > + rd->avg_nr_running = SCHED_CAPACITY_SCALE; > return 0; > > free_cpudl: > Based on the suggestions from RFC patch 5, I independently developed a patch. Verification showed highly satisfactory results: performance improved in the vast majority of tests. The test comparison is provided below: type pct hackbench +12.2% tbench +7.9% schbench 99th +2.6% stream-10 +38.5% stream-100 +52.0% unixbench +2.1% overall average +19.2% hackbench: group base patch pct -g 1 0.262 0.232 +11.5% -g 2 0.288 0.299 -4.0% -g 4 0.407 0.353 +13.3% -g 8 0.567 0.467 +17.8% -g 16 0.836 0.667 +20.2% -g 32 1.411 1.164 +17.5% -g 64 2.473 2.250 +9.0% tbench(throughput): clients base patch pct 1 475.97 487.16 +2.4% 2 714.48 757.36 +6.0% 4 1349.73 1484.51 +10.0% 8 2059.00 2193.00 +6.5% 16 4364.25 4305.11 -1.4% 32 8325.17 8102.79 -2.7% 64 12614.20 15140.07 +20.0% 128 15845.07 21614.20 +36.4% 256 20837.93 23567.93 +13.1% 512 29820.67 28243.37 -5.3% 1024 25969.93 26559.50 +2.3% schbench(99th): threads base patch pct 1 14.2 12.0 +15.5% 2 16.6 16.8 −1.2% 4 21.0 20.6 +1.9% 8 33.4 33.4 0.0% 16 54.6 50.4 +7.7% 32 93.4 79.0 +15.4% 64 196.4 219.6 −11.8% 128 4753.6 4804.8 −1.1% 256 7900.8 7822.4 +1.0% 512 11177.6 11318.4 −1.3% stream-10: operation base patch pct Copy 126408 188733 +49.3% Scale 124057 184036 +48.4% Add 160037 203560 +27.2% Triad 156224 201937 +29.3% stream-100: operation base patch pct Copy 172068 275389 +60.1% Scale 156385 248252 +58.7% Add 158850 231243 +45.6% Triad 162271 232983 +43.6% unixbench: cpu base patch pct 1 1808.3 1847.5 +2.2% 384 113834.2 116173.0 +2.1% -- thanks, kunyu