From nobody Mon Feb 9 09:42:34 2026 Received: from mail-wm1-f43.google.com (mail-wm1-f43.google.com [209.85.128.43]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id AC8A38F44 for ; Mon, 5 Feb 2024 02:11:39 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.128.43 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1707099101; cv=none; b=Wa+Lho0ITo/CH/OCQMo1VqXZj9TRELfkS5zOpFVXocRMT84RCaV5CjpxllJJbEOremanmUNVyX9qMvbGGVjUlhYR5QpGW3csgjxvDRtfZgG6c8aeGMLAV5Ahv7mFmBrvBw+GPEJ3Z4Cj9FMPOgJQ9YEPQyllhTEa/tlO8M4bSjE= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1707099101; c=relaxed/simple; bh=GmmdOiK7gVSr1LuIyV2fgmVs0Usd923RUm0wZ55KmaQ=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=fF0eBKqyJ292nqDhBodBCnzw4OmjFTAFUcPqw0Y7/riGd28ue0qqjzKESRCMsfx8KoZ7TYEMo9MawBGJCqh3JTDgd79JKBBb0vS0AZiR6hVqme+S+o8t74oAQ63E9Mz8VszX3pF9mOmqRTwUBtVqfLqvhrqkb9q8foIHUl+Yytg= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=layalina.io; spf=pass smtp.mailfrom=layalina.io; dkim=pass (2048-bit key) header.d=layalina-io.20230601.gappssmtp.com header.i=@layalina-io.20230601.gappssmtp.com header.b=HGwYy25D; arc=none smtp.client-ip=209.85.128.43 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=layalina.io Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=layalina.io Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=layalina-io.20230601.gappssmtp.com header.i=@layalina-io.20230601.gappssmtp.com header.b="HGwYy25D" Received: by mail-wm1-f43.google.com with SMTP id 5b1f17b1804b1-40fc549ab9bso23183735e9.0 for ; Sun, 04 Feb 2024 18:11:39 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=layalina-io.20230601.gappssmtp.com; s=20230601; t=1707099098; x=1707703898; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=V9mX0v9j5yCCNZk00rdpSpL3Xv1Bq3f+smlHRsHH9HE=; b=HGwYy25DvHfrDcryE6K//W4sL0nIrHqKCpRPo9nEwuQSpdkDQ/6gliFCNS1pHvud6+ DR4Np01BHdgVEiq6sY1qDIDDAWhuLdv9prvDtJrvx9i4AEf156viiMB96ph8TR1qUhzW imx6TWs4PEZSezCn0Y2xtuuLJjpFeLLHWAKDkIjIsT2ye1a1bCa1TsJa4yzBKvDA9vjq TabhyD1oXRbgfJhtiz5DpX85VabyWErM2B+S5qtsioNK5j8LHMIKglTWPYQhdWveugXH wplwv375P7JRdeBGAIOPWp3xx5WTuywk36WcIm/WOsjiLgxmDOgLp5BjYjCLtaf1hNnk xX3w== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1707099098; x=1707703898; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=V9mX0v9j5yCCNZk00rdpSpL3Xv1Bq3f+smlHRsHH9HE=; b=NUF2L/YNYtFq/rWgyMGnykMN5XyCl1ZMd4ZryOmnb2fTJbrsOW0G6eATGEzHfX7jmF N7Eo20obhNNlSZ1b+CpiyQXrU+C+txFQ5KiQ+xuPLUMFr866NRWsAuVbc9QeT/vAHu2v s7krhEg7+rsv3OSVEqnEgqhoXIpYQFLrsMzQZgAd3jbjMeaoQ63QVGuuKAwno9HgbQfU wBdgC56z8958CYn/s5hQIc1f+LhmqVtrUMxbPDoSY5NybE6caAFZXYv/RStZOkAhVolF cduGumISSgvWVOIWLnk21Zl9YoezsDnf9PPHHxgC60kfY/5mHyI8MTJLQ7Na4dZcRkx4 rJSw== X-Gm-Message-State: AOJu0Yxthzu9XBdRdMQvVdvOa2x7Ec58D9hF6EDH1xpqLxBJH6a2DhUI obaHePGXegIJveF++2khey64sd32HInHQ2GUJ4jKQPtZT/06N9QXsHyKp98urVXfdWt5Id8KWYp N X-Google-Smtp-Source: AGHT+IFqNIdlSbia0S9hnXbjw+Vmh40AmzDM1ejFONviqhcvAAIOH7xLfmxI2xMTA/8kjzhZCdMsbg== X-Received: by 2002:a05:600c:4750:b0:40f:cce4:38dd with SMTP id w16-20020a05600c475000b0040fcce438ddmr3953703wmo.20.1707099097808; Sun, 04 Feb 2024 18:11:37 -0800 (PST) X-Forwarded-Encrypted: i=0; AJvYcCVsqRFzOugBcUFYXU1cgdMRhqVuCuZ5zsy8WuswXaNaupOAIqfQSC9wwFk8QWFNcCdsuK9tcWFgv7oWAETNYDU6C3g+ErEmGqH+8U+shhih72Pfhc3YPLoMBW2wnwNwZfrBPZHWd0yG5EdyqLOz6/c738Osh/UeGzj1wQCeRKeMwJvc0F8jZCgU9kHBOqWiO9DsdYxbADgCrK6kFr3QCa4YSxLBMNYyBQ== Received: from airbuntu.. (host109-154-238-234.range109-154.btcentralplus.com. [109.154.238.234]) by smtp.gmail.com with ESMTPSA id j4-20020a5d4524000000b0033addeabd6fsm6999800wra.18.2024.02.04.18.11.37 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Sun, 04 Feb 2024 18:11:37 -0800 (PST) From: Qais Yousef To: Ingo Molnar , Peter Zijlstra , Vincent Guittot , Dietmar Eggemann Cc: linux-kernel@vger.kernel.org, "Pierre Gondois" , Qais Yousef Subject: [PATCH v5 1/2] sched/topology: Export asym_capacity_list Date: Mon, 5 Feb 2024 02:11:22 +0000 Message-Id: <20240205021123.2225933-2-qyousef@layalina.io> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20240205021123.2225933-1-qyousef@layalina.io> References: <20240205021123.2225933-1-qyousef@layalina.io> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" So that we can use it to iterate through available capacities in the system. Sort asym_cap_list in descending order as expected users are likely to be interested on the highest capacity first. Make the list RCU protected to allow for cheap access in hot paths. Signed-off-by: Qais Yousef --- kernel/sched/sched.h | 14 ++++++++++++++ kernel/sched/topology.c | 43 ++++++++++++++++++++++++----------------- 2 files changed, 39 insertions(+), 18 deletions(-) diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index e58a54bda77d..a653017a1b9b 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -109,6 +109,20 @@ extern int sysctl_sched_rt_period; extern int sysctl_sched_rt_runtime; extern int sched_rr_timeslice; =20 +/* + * Asymmetric CPU capacity bits + */ +struct asym_cap_data { + struct list_head link; + struct rcu_head rcu; + unsigned long capacity; + unsigned long cpus[]; +}; + +extern struct list_head asym_cap_list; + +#define cpu_capacity_span(asym_data) to_cpumask((asym_data)->cpus) + /* * Helpers for converting nanosecond timing to jiffy resolution */ diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c index 10d1391e7416..1505677e4247 100644 --- a/kernel/sched/topology.c +++ b/kernel/sched/topology.c @@ -1329,24 +1329,13 @@ static void init_sched_groups_capacity(int cpu, str= uct sched_domain *sd) update_group_capacity(sd, cpu); } =20 -/* - * Asymmetric CPU capacity bits - */ -struct asym_cap_data { - struct list_head link; - unsigned long capacity; - unsigned long cpus[]; -}; - /* * Set of available CPUs grouped by their corresponding capacities * Each list entry contains a CPU mask reflecting CPUs that share the same * capacity. * The lifespan of data is unlimited. */ -static LIST_HEAD(asym_cap_list); - -#define cpu_capacity_span(asym_data) to_cpumask((asym_data)->cpus) +LIST_HEAD(asym_cap_list); =20 /* * Verify whether there is any CPU capacity asymmetry in a given sched dom= ain. @@ -1386,21 +1375,39 @@ asym_cpu_capacity_classify(const struct cpumask *sd= _span, =20 } =20 +static void free_asym_cap_entry(struct rcu_head *head) +{ + struct asym_cap_data *entry =3D container_of(head, struct asym_cap_data, = rcu); + kfree(entry); +} + static inline void asym_cpu_capacity_update_data(int cpu) { unsigned long capacity =3D arch_scale_cpu_capacity(cpu); - struct asym_cap_data *entry =3D NULL; + struct asym_cap_data *insert_entry =3D NULL; + struct asym_cap_data *entry; =20 + /* + * Search if capacity already exits. If not, track which the entry + * where we should insert to keep the list ordered descendingly. + */ list_for_each_entry(entry, &asym_cap_list, link) { if (capacity =3D=3D entry->capacity) goto done; + else if (!insert_entry && capacity > entry->capacity) + insert_entry =3D list_prev_entry(entry, link); } =20 entry =3D kzalloc(sizeof(*entry) + cpumask_size(), GFP_KERNEL); if (WARN_ONCE(!entry, "Failed to allocate memory for asymmetry data\n")) return; entry->capacity =3D capacity; - list_add(&entry->link, &asym_cap_list); + + /* If NULL then the new capacity is the smallest, add last. */ + if (!insert_entry) + list_add_tail_rcu(&entry->link, &asym_cap_list); + else + list_add_rcu(&entry->link, &insert_entry->link); done: __cpumask_set_cpu(cpu, cpu_capacity_span(entry)); } @@ -1423,8 +1430,8 @@ static void asym_cpu_capacity_scan(void) =20 list_for_each_entry_safe(entry, next, &asym_cap_list, link) { if (cpumask_empty(cpu_capacity_span(entry))) { - list_del(&entry->link); - kfree(entry); + list_del_rcu(&entry->link); + call_rcu(&entry->rcu, free_asym_cap_entry); } } =20 @@ -1434,8 +1441,8 @@ static void asym_cpu_capacity_scan(void) */ if (list_is_singular(&asym_cap_list)) { entry =3D list_first_entry(&asym_cap_list, typeof(*entry), link); - list_del(&entry->link); - kfree(entry); + list_del_rcu(&entry->link); + call_rcu(&entry->rcu, free_asym_cap_entry); } } =20 --=20 2.34.1 From nobody Mon Feb 9 09:42:34 2026 Received: from mail-wm1-f48.google.com (mail-wm1-f48.google.com [209.85.128.48]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 51B1F8F47 for ; Mon, 5 Feb 2024 02:11:40 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.128.48 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1707099102; cv=none; b=p0xiPgMzDd6xO90Q9Mixgscl61yLBv9ZGRNc81kmdWI3wLjP9V+XZz+98JNJglaOgh8VuntJB9+qKT84f50ZDwS1ZnckHsLELsOg5YIMF4s23YA1OpVB4i1klF4LD5NgV2Y5foc48zuZpLRySOCf/3gfAvn3dBnVo3Y3Zw6DZDI= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1707099102; c=relaxed/simple; bh=PMsc4K9BD1uYrMwqzPyh9tQ6u4MAjJ4W+8DnzZ9warA=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=CNWGafpZ1VaFh0orHN/QyILj6Uo7G5UWAIXB2Hb7Mb+GFrvREZfLQnNmreutnCAKG2419KM64ryND1z60rCcWPtabZhn/3O+6OHm556jWKXsj3diiJbQ9161fTx4n9E0TYYZOcH0kcfKwOjwLxjZKEN+m7rt6JiDkOZz/2zK4CQ= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=layalina.io; spf=pass smtp.mailfrom=layalina.io; dkim=pass (2048-bit key) header.d=layalina-io.20230601.gappssmtp.com header.i=@layalina-io.20230601.gappssmtp.com header.b=xktyR0bY; arc=none smtp.client-ip=209.85.128.48 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=layalina.io Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=layalina.io Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=layalina-io.20230601.gappssmtp.com header.i=@layalina-io.20230601.gappssmtp.com header.b="xktyR0bY" Received: by mail-wm1-f48.google.com with SMTP id 5b1f17b1804b1-40fdd5bbe65so429705e9.2 for ; Sun, 04 Feb 2024 18:11:40 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=layalina-io.20230601.gappssmtp.com; s=20230601; t=1707099098; x=1707703898; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=hrqO2B+XixMyeCid6/wthHwgVgi14v6y2AEDv2YkXgI=; b=xktyR0bYJ+z90fR8rjTCh+jByWk6K/SoRNyeDIqQdchtofwZf5DFkJq7As8OvZQiGo FCvKr5xP1ptpcxtosejL0s4+iovnIGltgdSWt5nPgL74UeFDUjqbFk1SoMLoSVTxui1P Zc9K3NjyrRTRB4Ln1N/kSf1xQsn28n0GYbe4WDVpyRc/7JiHWgAO15G5KXpzW+AF9mAf pKfdHuMJ7Bq7jq/ZOOfE8fxfkRP+79hAUwu5qRbZDxy/mH5FODZZmOOXPo31zndK6ylf 5E1Gl4JDGLSMGyvzhgeshsZ02+8FHq37Knr8MKr/rurgEIdLMcW6d/dGtPvP3Vrd5JUZ UcoQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1707099098; x=1707703898; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=hrqO2B+XixMyeCid6/wthHwgVgi14v6y2AEDv2YkXgI=; b=j7BDZyJ9LCRyLKgcTWh0ZjDkc/yZ/9B7c08dyfoCBh4bqYRus72VGUYbXXHHTMAz86 BMjb/u6TnDm27OdZ8ojlnHksLCMQ2myNl3PT6QoogvJdi4IpZ77wEfO3T8UXi0V/6GvS rfct3yJKK1tcOR1erD6JgIP5Xd9QL5KihhcyU0vXjJvxK14xmHSeqxn74unhBVU0f/6x 8LwpYuLrlzadQ4w69okQQphJQoLqCSAT/KqTGl8nhHCOcGwEtiPQIdyuVfO7i09rw09D HuKmvG0OPUdk8td/OFMG5sVGDkuYjGTmWE4mCM/pjmzlPfp2QDXjmtJ2BJfdSViCh8pb ShFg== X-Gm-Message-State: AOJu0YxLcCNZEEukYS7JB/+xUPzAero09hSlhTfd9Q+2tlEnk5XGr4K+ CY6/yPlGGFVWg9Hf5kPhJlilkx4JmZqWTtfZULu6UiEeQcGT+mkXZhQCC7Q5f1c= X-Google-Smtp-Source: AGHT+IGyVxX6aou1hcQpNo4ivtaHKilGSv1jF1i64t31J9mV465YpHZX7iyoJJlRJUtobD/ZFvYxzQ== X-Received: by 2002:adf:f8c6:0:b0:33a:ea26:6a91 with SMTP id f6-20020adff8c6000000b0033aea266a91mr6839013wrq.0.1707099098627; Sun, 04 Feb 2024 18:11:38 -0800 (PST) X-Forwarded-Encrypted: i=0; AJvYcCX6fU1gWVTH1AXCLzFeKYNKzOP9Ot7X+FrcWowicAYHIMDaauqeiixn6UmgcA4+k0HPoPJz2NAtff/m/JHIa6+Y6RVpRgOqgUWmXdl2kfuO3xKMC98/1qK/597jzpQk/csCNOxFwyWueu05XEiNtvp9L/oTt5jupHx6EOqzP4199K9atFHotZEEHjwEpHiSm590l0zMr8pfOME83uCyZRhWy340xyp+kQ== Received: from airbuntu.. (host109-154-238-234.range109-154.btcentralplus.com. [109.154.238.234]) by smtp.gmail.com with ESMTPSA id j4-20020a5d4524000000b0033addeabd6fsm6999800wra.18.2024.02.04.18.11.37 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Sun, 04 Feb 2024 18:11:38 -0800 (PST) From: Qais Yousef To: Ingo Molnar , Peter Zijlstra , Vincent Guittot , Dietmar Eggemann Cc: linux-kernel@vger.kernel.org, "Pierre Gondois" , Qais Yousef Subject: [PATCH v5 2/2] sched/fair: Check a task has a fitting cpu when updating misfit Date: Mon, 5 Feb 2024 02:11:23 +0000 Message-Id: <20240205021123.2225933-3-qyousef@layalina.io> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20240205021123.2225933-1-qyousef@layalina.io> References: <20240205021123.2225933-1-qyousef@layalina.io> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" If a misfit task is affined to a subset of the possible cpus, we need to verify that one of these cpus can fit it. Otherwise the load balancer code will continuously trigger needlessly leading the balance_interval to increase in return and eventually end up with a situation where real imbalances take a long time to address because of this impossible imbalance situation. This can happen in Android world where it's common for background tasks to be restricted to little cores. Similarly if we can't fit the biggest core, triggering misfit is pointless as it is the best we can ever get on this system. To be able to detect that; we use asym_cap_list to iterate through capacities in the system to see if the task is able to run at a higher capacity level based on its p->cpus_ptr. We do that when the affinity change, a fair task is forked, or when a task switched to fair policy. We store the max_allowed_capacity in task_struct to allow for cheap comparison in the fast path. Improve check_misfit_status() function to be more readable. At one iteration of the patch it was thought we can drop one of the checks. The current form hopefully should make it more obvious what is being checked and why. Test: =3D=3D=3D=3D=3D Add trace_printk("balance_interval =3D %lu\n", interval) in get_sd_balance_interval(). run if [ "$MASK" !=3D "0" ]; then adb shell "taskset -a $MASK cat /dev/zero > /dev/null" fi sleep 10 // parse ftrace buffer counting the occurrence of each valaue Where MASK is either: * 0: no busy task running * 1: busy task is pinned to 1 cpu; handled today to not cause misfit * f: busy task pinned to little cores, simulates busy background task, demonstrates the problem to be fixed Results: =3D=3D=3D=3D=3D=3D=3D=3D Note how occurrence of balance_interval =3D 128 overshoots for MASK =3D f. BEFORE ------ MASK=3D0 1 balance_interval =3D 175 120 balance_interval =3D 128 846 balance_interval =3D 64 55 balance_interval =3D 63 215 balance_interval =3D 32 2 balance_interval =3D 31 2 balance_interval =3D 16 4 balance_interval =3D 8 1870 balance_interval =3D 4 65 balance_interval =3D 2 MASK=3D1 27 balance_interval =3D 175 37 balance_interval =3D 127 840 balance_interval =3D 64 167 balance_interval =3D 63 449 balance_interval =3D 32 84 balance_interval =3D 31 304 balance_interval =3D 16 1156 balance_interval =3D 8 2781 balance_interval =3D 4 428 balance_interval =3D 2 MASK=3Df 1 balance_interval =3D 175 1328 balance_interval =3D 128 44 balance_interval =3D 64 101 balance_interval =3D 63 25 balance_interval =3D 32 5 balance_interval =3D 31 23 balance_interval =3D 16 23 balance_interval =3D 8 4306 balance_interval =3D 4 177 balance_interval =3D 2 AFTER ----- Note how the high values almost disappear for all MASK values. The system has background tasks that could trigger the problem without simulate it even with MASK=3D0. MASK=3D0 103 balance_interval =3D 63 19 balance_interval =3D 31 194 balance_interval =3D 8 4827 balance_interval =3D 4 179 balance_interval =3D 2 MASK=3D1 131 balance_interval =3D 63 1 balance_interval =3D 31 87 balance_interval =3D 8 3600 balance_interval =3D 4 7 balance_interval =3D 2 MASK=3Df 8 balance_interval =3D 127 182 balance_interval =3D 63 3 balance_interval =3D 31 9 balance_interval =3D 16 415 balance_interval =3D 8 3415 balance_interval =3D 4 21 balance_interval =3D 2 Signed-off-by: Qais Yousef --- include/linux/sched.h | 1 + init/init_task.c | 1 + kernel/sched/fair.c | 84 +++++++++++++++++++++++++++++++++---------- 3 files changed, 68 insertions(+), 18 deletions(-) diff --git a/include/linux/sched.h b/include/linux/sched.h index 03bfe9ab2951..1e7bf52de607 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -827,6 +827,7 @@ struct task_struct { #endif =20 unsigned int policy; + unsigned long max_allowed_capacity; int nr_cpus_allowed; const cpumask_t *cpus_ptr; cpumask_t *user_cpus_ptr; diff --git a/init/init_task.c b/init/init_task.c index 5727d42149c3..01b3199d4cde 100644 --- a/init/init_task.c +++ b/init/init_task.c @@ -82,6 +82,7 @@ struct task_struct init_task .cpus_ptr =3D &init_task.cpus_mask, .user_cpus_ptr =3D NULL, .cpus_mask =3D CPU_MASK_ALL, + .max_allowed_capacity =3D SCHED_CAPACITY_SCALE, .nr_cpus_allowed=3D NR_CPUS, .mm =3D NULL, .active_mm =3D &init_mm, diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index b803030c3a03..8b8035f5c8f6 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -5092,24 +5092,36 @@ static inline int task_fits_cpu(struct task_struct = *p, int cpu) =20 static inline void update_misfit_status(struct task_struct *p, struct rq *= rq) { + unsigned long cpu_cap; + int cpu =3D cpu_of(rq); + if (!sched_asym_cpucap_active()) return; =20 - if (!p || p->nr_cpus_allowed =3D=3D 1) { - rq->misfit_task_load =3D 0; - return; - } + if (!p || p->nr_cpus_allowed =3D=3D 1) + goto out; =20 - if (task_fits_cpu(p, cpu_of(rq))) { - rq->misfit_task_load =3D 0; - return; - } + cpu_cap =3D arch_scale_cpu_capacity(cpu); + + /* If we can't fit the biggest CPU, that's the best we can ever get. */ + if (cpu_cap =3D=3D rq->rd->max_cpu_capacity) + goto out; + + /* Affinity allows us to go somewhere higher? */ + if (cpu_cap =3D=3D p->max_allowed_capacity) + goto out; + + if (task_fits_cpu(p, cpu)) + goto out; =20 /* * Make sure that misfit_task_load will not be null even if * task_h_load() returns 0. */ rq->misfit_task_load =3D max_t(unsigned long, task_h_load(p), 1); + return; +out: + rq->misfit_task_load =3D 0; } =20 #else /* CONFIG_SMP */ @@ -8241,6 +8253,36 @@ static void task_dead_fair(struct task_struct *p) remove_entity_load_avg(&p->se); } =20 +/* + * Check the max capacity the task is allowed to run at for misfit detecti= on. + */ +static void set_task_max_allowed_capacity(struct task_struct *p) +{ + struct asym_cap_data *entry; + + if (!sched_asym_cpucap_active()) + return; + + rcu_read_lock(); + list_for_each_entry_rcu(entry, &asym_cap_list, link) { + cpumask_t *cpumask; + + cpumask =3D cpu_capacity_span(entry); + if (!cpumask_intersects(p->cpus_ptr, cpumask)) + continue; + + p->max_allowed_capacity =3D entry->capacity; + break; + } + rcu_read_unlock(); +} + +static void set_cpus_allowed_fair(struct task_struct *p, struct affinity_c= ontext *ctx) +{ + set_cpus_allowed_common(p, ctx); + set_task_max_allowed_capacity(p); +} + static int balance_fair(struct rq *rq, struct task_struct *prev, struct rq_flags *rf) { @@ -9601,16 +9643,18 @@ check_cpu_capacity(struct rq *rq, struct sched_doma= in *sd) (arch_scale_cpu_capacity(cpu_of(rq)) * 100)); } =20 -/* - * Check whether a rq has a misfit task and if it looks like we can actual= ly - * help that task: we can migrate the task to a CPU of higher capacity, or - * the task's current CPU is heavily pressured. - */ -static inline int check_misfit_status(struct rq *rq, struct sched_domain *= sd) +/* Check if the rq has a misfit task */ +static inline bool check_misfit_status(struct rq *rq, struct sched_domain = *sd) { - return rq->misfit_task_load && - (arch_scale_cpu_capacity(rq->cpu) < rq->rd->max_cpu_capacity || - check_cpu_capacity(rq, sd)); + if (!rq->misfit_task_load) + return false; + + /* Can we migrate to a CPU with higher capacity? */ + if (arch_scale_cpu_capacity(rq->cpu) < rq->rd->max_cpu_capacity) + return true; + + /* Is the task's CPU being heavily pressured? */ + return check_cpu_capacity(rq, sd); } =20 /* @@ -12647,6 +12691,8 @@ static void task_fork_fair(struct task_struct *p) rq_lock(rq, &rf); update_rq_clock(rq); =20 + set_task_max_allowed_capacity(p); + cfs_rq =3D task_cfs_rq(current); curr =3D cfs_rq->curr; if (curr) @@ -12770,6 +12816,8 @@ static void switched_to_fair(struct rq *rq, struct = task_struct *p) { attach_task_cfs_rq(p); =20 + set_task_max_allowed_capacity(p); + if (task_on_rq_queued(p)) { /* * We were most likely switched from sched_rt, so @@ -13154,7 +13202,7 @@ DEFINE_SCHED_CLASS(fair) =3D { .rq_offline =3D rq_offline_fair, =20 .task_dead =3D task_dead_fair, - .set_cpus_allowed =3D set_cpus_allowed_common, + .set_cpus_allowed =3D set_cpus_allowed_fair, #endif =20 .task_tick =3D task_tick_fair, --=20 2.34.1