From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-15.8 required=3.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id D3930C433F5 for ; Thu, 2 Sep 2021 21:55:03 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 81268610D2 for ; Thu, 2 Sep 2021 21:55:03 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.4.1 mail.kernel.org 81268610D2 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=linux-foundation.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=kvack.org Received: by kanga.kvack.org (Postfix) id 211936B00E5; Thu, 2 Sep 2021 17:55:03 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id 1C01A6B00E6; Thu, 2 Sep 2021 17:55:03 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 089078D0001; Thu, 2 Sep 2021 17:55:03 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0239.hostedemail.com [216.40.44.239]) by kanga.kvack.org (Postfix) with ESMTP id ECB946B00E5 for ; Thu, 2 Sep 2021 17:55:02 -0400 (EDT) Received: from smtpin08.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay01.hostedemail.com (Postfix) with ESMTP id A809A1840B39F for ; Thu, 2 Sep 2021 21:55:02 +0000 (UTC) X-FDA: 78543989244.08.7E73159 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by imf08.hostedemail.com (Postfix) with ESMTP id 5A47130000A8 for ; Thu, 2 Sep 2021 21:55:02 +0000 (UTC) Received: by mail.kernel.org (Postfix) with ESMTPSA id 455D3610E5; Thu, 2 Sep 2021 21:55:01 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linux-foundation.org; s=korg; t=1630619701; bh=2bWkvo5qMGLmFr37K5saCrpJlhtv6vRpGyKU1GQHpTg=; h=Date:From:To:Subject:In-Reply-To:From; b=ZqpshUUOa9HIPpo+H0gZ8AHhI3PxUo4Z4kNs2GdW7QWFkkI4AExVdTsj4os2Vl0Yh 2FtMUgNTlTw6DgrSsza5jeoVPHqRPFDx8FJe7Bzdk9Z0Uf9N1Brh03GgaePOVipJfv WnkFDd4zhcgz+6k7+xDrSoTWIv/eq2Kj91lbMdcE= Date: Thu, 02 Sep 2021 14:55:00 -0700 From: Andrew Morton To: akpm@linux-foundation.org, guro@fb.com, hannes@cmpxchg.org, hdanton@sina.com, linux-mm@kvack.org, mhocko@kernel.org, mkoutny@suse.com, mm-commits@vger.kernel.org, shakeelb@google.com, songmuchun@bytedance.com, tj@kernel.org, torvalds@linux-foundation.org, ying.huang@intel.com Subject: [patch 096/212] memcg: switch lruvec stats to rstat Message-ID: <20210902215500.rnJobPz5l%akpm@linux-foundation.org> In-Reply-To: <20210902144820.78957dff93d7bea620d55a89@linux-foundation.org> User-Agent: s-nail v14.8.16 MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Authentication-Results: imf08.hostedemail.com; dkim=pass header.d=linux-foundation.org header.s=korg header.b=ZqpshUUO; spf=pass (imf08.hostedemail.com: domain of akpm@linux-foundation.org designates 198.145.29.99 as permitted sender) smtp.mailfrom=akpm@linux-foundation.org; dmarc=none X-Stat-Signature: xmztjhrb4t7us7h3d11rz7gg3rrwwumw X-Rspamd-Queue-Id: 5A47130000A8 X-Rspamd-Server: rspam04 X-HE-Tag: 1630619702-636747 X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: =46rom: Shakeel Butt Subject: memcg: switch lruvec stats to rstat The commit 2d146aa3aa84 ("mm: memcontrol: switch to rstat") switched memcg stats to rstat infrastructure but skipped the conversion of the lruvec stats as such stats are read in the performance critical code paths and flushing stats may have impacted the performances of the applications.=20 This patch converts the lruvec stats to rstat and later patches add mechanisms to keep the performance impact to minimum. The rstat conversion comes with the price i.e. memory cost. Effectively this patch reverts the savings done by the commit f3344adf38bd ("mm: memcontrol: optimize per-lruvec stats counter memory usage"). However this cost is justified due to negative impact of the inaccurate lruvec stats on many heuristics. One such case is reported in [1]. The memory reclaim code is filled with plethora of heuristics and many of those heuristics reads the lruvec stats. So, inaccurate stats can make such heuristics ineffective. [1] reports the impact of inaccurate lruvec stats on the "cache trim mode" heuristic. Inaccurate lruvec stats can impact the deactivation and aging anon heuristics as well. [1] https://lore.kernel.org/linux-mm/20210311004449.1170308-1-ying.huang@in= tel.com/ Link: https://lkml.kernel.org/r/20210716212137.1391164-1-shakeelb@google.com Link: https://lkml.kernel.org/r/20210714013948.270662-1-shakeelb@google.com Signed-off-by: Shakeel Butt Cc: Tejun Heo Cc: Johannes Weiner Cc: Muchun Song Cc: Michal Hocko Cc: Roman Gushchin Cc: Huang Ying Cc: Hillf Danton Cc: Michal Koutn=C3=BD Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 42 ++++++------ mm/memcontrol.c | 114 +++++++++++------------------------ 2 files changed, 58 insertions(+), 98 deletions(-) --- a/include/linux/memcontrol.h~memcg-switch-lruvec-stats-to-rstat +++ a/include/linux/memcontrol.h @@ -105,14 +105,6 @@ struct mem_cgroup_reclaim_iter { unsigned int generation; }; =20 -struct lruvec_stat { - long count[NR_VM_NODE_STAT_ITEMS]; -}; - -struct batched_lruvec_stat { - s32 count[NR_VM_NODE_STAT_ITEMS]; -}; - /* * Bitmap and deferred work of shrinker::id corresponding to memcg-aware * shrinkers, which have elements charged to this memcg. @@ -123,24 +115,30 @@ struct shrinker_info { unsigned long *map; }; =20 +struct lruvec_stats_percpu { + /* Local (CPU and cgroup) state */ + long state[NR_VM_NODE_STAT_ITEMS]; + + /* Delta calculation for lockless upward propagation */ + long state_prev[NR_VM_NODE_STAT_ITEMS]; +}; + +struct lruvec_stats { + /* Aggregated (CPU and subtree) state */ + long state[NR_VM_NODE_STAT_ITEMS]; + + /* Pending child counts during tree propagation */ + long state_pending[NR_VM_NODE_STAT_ITEMS]; +}; + /* * per-node information in memory controller. */ struct mem_cgroup_per_node { struct lruvec lruvec; =20 - /* - * Legacy local VM stats. This should be struct lruvec_stat and - * cannot be optimized to struct batched_lruvec_stat. Because - * the threshold of the lruvec_stat_cpu can be as big as - * MEMCG_CHARGE_BATCH * PAGE_SIZE. It can fit into s32. But this - * filed has no upper limit. - */ - struct lruvec_stat __percpu *lruvec_stat_local; - - /* Subtree VM stats (batched updates) */ - struct batched_lruvec_stat __percpu *lruvec_stat_cpu; - atomic_long_t lruvec_stat[NR_VM_NODE_STAT_ITEMS]; + struct lruvec_stats_percpu __percpu *lruvec_stats_percpu; + struct lruvec_stats lruvec_stats; =20 unsigned long lru_zone_size[MAX_NR_ZONES][NR_LRU_LISTS]; =20 @@ -997,7 +995,7 @@ static inline unsigned long lruvec_page_ return node_page_state(lruvec_pgdat(lruvec), idx); =20 pn =3D container_of(lruvec, struct mem_cgroup_per_node, lruvec); - x =3D atomic_long_read(&pn->lruvec_stat[idx]); + x =3D READ_ONCE(pn->lruvec_stats.state[idx]); #ifdef CONFIG_SMP if (x < 0) x =3D 0; @@ -1017,7 +1015,7 @@ static inline unsigned long lruvec_page_ =20 pn =3D container_of(lruvec, struct mem_cgroup_per_node, lruvec); for_each_possible_cpu(cpu) - x +=3D per_cpu(pn->lruvec_stat_local->count[idx], cpu); + x +=3D per_cpu(pn->lruvec_stats_percpu->state[idx], cpu); #ifdef CONFIG_SMP if (x < 0) x =3D 0; --- a/mm/memcontrol.c~memcg-switch-lruvec-stats-to-rstat +++ a/mm/memcontrol.c @@ -660,23 +660,11 @@ static unsigned long memcg_page_state_lo return x; } =20 -static struct mem_cgroup_per_node * -parent_nodeinfo(struct mem_cgroup_per_node *pn, int nid) -{ - struct mem_cgroup *parent; - - parent =3D parent_mem_cgroup(pn->memcg); - if (!parent) - return NULL; - return parent->nodeinfo[nid]; -} - void __mod_memcg_lruvec_state(struct lruvec *lruvec, enum node_stat_item i= dx, int val) { struct mem_cgroup_per_node *pn; struct mem_cgroup *memcg; - long x, threshold =3D MEMCG_CHARGE_BATCH; =20 pn =3D container_of(lruvec, struct mem_cgroup_per_node, lruvec); memcg =3D pn->memcg; @@ -685,21 +673,7 @@ void __mod_memcg_lruvec_state(struct lru __mod_memcg_state(memcg, idx, val); =20 /* Update lruvec */ - __this_cpu_add(pn->lruvec_stat_local->count[idx], val); - - if (vmstat_item_in_bytes(idx)) - threshold <<=3D PAGE_SHIFT; - - x =3D val + __this_cpu_read(pn->lruvec_stat_cpu->count[idx]); - if (unlikely(abs(x) > threshold)) { - pg_data_t *pgdat =3D lruvec_pgdat(lruvec); - struct mem_cgroup_per_node *pi; - - for (pi =3D pn; pi; pi =3D parent_nodeinfo(pi, pgdat->node_id)) - atomic_long_add(x, &pi->lruvec_stat[idx]); - x =3D 0; - } - __this_cpu_write(pn->lruvec_stat_cpu->count[idx], x); + __this_cpu_add(pn->lruvec_stats_percpu->state[idx], val); } =20 /** @@ -2278,40 +2252,13 @@ static void drain_all_stock(struct mem_c mutex_unlock(&percpu_charge_mutex); } =20 -static void memcg_flush_lruvec_page_state(struct mem_cgroup *memcg, int cp= u) -{ - int nid; - - for_each_node(nid) { - struct mem_cgroup_per_node *pn =3D memcg->nodeinfo[nid]; - unsigned long stat[NR_VM_NODE_STAT_ITEMS]; - struct batched_lruvec_stat *lstatc; - int i; - - lstatc =3D per_cpu_ptr(pn->lruvec_stat_cpu, cpu); - for (i =3D 0; i < NR_VM_NODE_STAT_ITEMS; i++) { - stat[i] =3D lstatc->count[i]; - lstatc->count[i] =3D 0; - } - - do { - for (i =3D 0; i < NR_VM_NODE_STAT_ITEMS; i++) - atomic_long_add(stat[i], &pn->lruvec_stat[i]); - } while ((pn =3D parent_nodeinfo(pn, nid))); - } -} - static int memcg_hotplug_cpu_dead(unsigned int cpu) { struct memcg_stock_pcp *stock; - struct mem_cgroup *memcg; =20 stock =3D &per_cpu(memcg_stock, cpu); drain_stock(stock); =20 - for_each_mem_cgroup(memcg) - memcg_flush_lruvec_page_state(memcg, cpu); - return 0; } =20 @@ -5118,17 +5065,9 @@ static int alloc_mem_cgroup_per_node_inf if (!pn) return 1; =20 - pn->lruvec_stat_local =3D alloc_percpu_gfp(struct lruvec_stat, - GFP_KERNEL_ACCOUNT); - if (!pn->lruvec_stat_local) { - kfree(pn); - return 1; - } - - pn->lruvec_stat_cpu =3D alloc_percpu_gfp(struct batched_lruvec_stat, - GFP_KERNEL_ACCOUNT); - if (!pn->lruvec_stat_cpu) { - free_percpu(pn->lruvec_stat_local); + pn->lruvec_stats_percpu =3D alloc_percpu_gfp(struct lruvec_stats_percpu, + GFP_KERNEL_ACCOUNT); + if (!pn->lruvec_stats_percpu) { kfree(pn); return 1; } @@ -5149,8 +5088,7 @@ static void free_mem_cgroup_per_node_inf if (!pn) return; =20 - free_percpu(pn->lruvec_stat_cpu); - free_percpu(pn->lruvec_stat_local); + free_percpu(pn->lruvec_stats_percpu); kfree(pn); } =20 @@ -5166,15 +5104,7 @@ static void __mem_cgroup_free(struct mem =20 static void mem_cgroup_free(struct mem_cgroup *memcg) { - int cpu; - memcg_wb_domain_exit(memcg); - /* - * Flush percpu lruvec stats to guarantee the value - * correctness on parent's and all ancestor levels. - */ - for_each_online_cpu(cpu) - memcg_flush_lruvec_page_state(memcg, cpu); __mem_cgroup_free(memcg); } =20 @@ -5407,7 +5337,7 @@ static void mem_cgroup_css_rstat_flush(s struct mem_cgroup *parent =3D parent_mem_cgroup(memcg); struct memcg_vmstats_percpu *statc; long delta, v; - int i; + int i, nid; =20 statc =3D per_cpu_ptr(memcg->vmstats_percpu, cpu); =20 @@ -5455,6 +5385,36 @@ static void mem_cgroup_css_rstat_flush(s if (parent) parent->vmstats.events_pending[i] +=3D delta; } + + for_each_node_state(nid, N_MEMORY) { + struct mem_cgroup_per_node *pn =3D memcg->nodeinfo[nid]; + struct mem_cgroup_per_node *ppn =3D NULL; + struct lruvec_stats_percpu *lstatc; + + if (parent) + ppn =3D parent->nodeinfo[nid]; + + lstatc =3D per_cpu_ptr(pn->lruvec_stats_percpu, cpu); + + for (i =3D 0; i < NR_VM_NODE_STAT_ITEMS; i++) { + delta =3D pn->lruvec_stats.state_pending[i]; + if (delta) + pn->lruvec_stats.state_pending[i] =3D 0; + + v =3D READ_ONCE(lstatc->state[i]); + if (v !=3D lstatc->state_prev[i]) { + delta +=3D v - lstatc->state_prev[i]; + lstatc->state_prev[i] =3D v; + } + + if (!delta) + continue; + + pn->lruvec_stats.state[i] +=3D delta; + if (ppn) + ppn->lruvec_stats.state_pending[i] +=3D delta; + } + } } =20 #ifdef CONFIG_MMU @@ -6388,6 +6348,8 @@ static int memory_numa_stat_show(struct int i; struct mem_cgroup *memcg =3D mem_cgroup_from_seq(m); =20 + cgroup_rstat_flush(memcg->css.cgroup); + for (i =3D 0; i < ARRAY_SIZE(memory_stats); i++) { int nid; =20 _