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=-15.3 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,NICE_REPLY_A,SPF_HELO_NONE,SPF_PASS,USER_AGENT_SANE_1 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 EAD8CC433DB for ; Fri, 29 Jan 2021 13:01:01 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 593D264DED for ; Fri, 29 Jan 2021 13:01:01 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 593D264DED Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.cz Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 876946B0005; Fri, 29 Jan 2021 08:01:00 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 7FF736B0006; Fri, 29 Jan 2021 08:01:00 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 6EE946B006C; Fri, 29 Jan 2021 08:01:00 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0139.hostedemail.com [216.40.44.139]) by kanga.kvack.org (Postfix) with ESMTP id 563156B0005 for ; Fri, 29 Jan 2021 08:01:00 -0500 (EST) Received: from smtpin10.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay04.hostedemail.com (Postfix) with ESMTP id 5880912C3 for ; Fri, 29 Jan 2021 13:00:57 +0000 (UTC) X-FDA: 77758822554.10.flag91_2100bf7275a8 Received: from filter.hostedemail.com (10.5.16.251.rfc1918.com [10.5.16.251]) by smtpin10.hostedemail.com (Postfix) with ESMTP id 7A7ED16A4D8 for ; Fri, 29 Jan 2021 13:00:54 +0000 (UTC) X-HE-Tag: flag91_2100bf7275a8 X-Filterd-Recvd-Size: 8677 Received: from mx2.suse.de (mx2.suse.de [195.135.220.15]) by imf45.hostedemail.com (Postfix) with ESMTP for ; Fri, 29 Jan 2021 13:00:53 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.221.27]) by mx2.suse.de (Postfix) with ESMTP id 12833AC41; Fri, 29 Jan 2021 13:00:52 +0000 (UTC) To: Yang Shi , guro@fb.com, ktkhai@virtuozzo.com, shakeelb@google.com, david@fromorbit.com, hannes@cmpxchg.org, mhocko@suse.com, akpm@linux-foundation.org Cc: linux-mm@kvack.org, linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org References: <20210127233345.339910-1-shy828301@gmail.com> <20210127233345.339910-8-shy828301@gmail.com> From: Vlastimil Babka Subject: Re: [v5 PATCH 07/11] mm: vmscan: add per memcg shrinker nr_deferred Message-ID: <6b0638ba-2513-67f5-8ef1-9e60a7d9ded6@suse.cz> Date: Fri, 29 Jan 2021 14:00:51 +0100 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.6.1 MIME-Version: 1.0 In-Reply-To: <20210127233345.339910-8-shy828301@gmail.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: quoted-printable 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: On 1/28/21 12:33 AM, Yang Shi wrote: > Currently the number of deferred objects are per shrinker, but some sla= bs, for example, > vfs inode/dentry cache are per memcg, this would result in poor isolati= on among memcgs. >=20 > The deferred objects typically are generated by __GFP_NOFS allocations,= one memcg with > excessive __GFP_NOFS allocations may blow up deferred objects, then oth= er innocent memcgs > may suffer from over shrink, excessive reclaim latency, etc. >=20 > For example, two workloads run in memcgA and memcgB respectively, workl= oad in B is vfs > heavy workload. Workload in A generates excessive deferred objects, th= en B's vfs cache > might be hit heavily (drop half of caches) by B's limit reclaim or glob= al reclaim. >=20 > We observed this hit in our production environment which was running vf= s heavy workload > shown as the below tracing log: >=20 > <...>-409454 [016] .... 28286961.747146: mm_shrink_slab_start: super_ca= che_scan+0x0/0x1a0 ffff9a83046f3458: > nid: 1 objects to shrink 3641681686040 gfp_flags GFP_HIGHUSER_MOVABLE|_= _GFP_ZERO pgs_scanned 1 lru_pgs 15721 > cache items 246404277 delta 31345 total_scan 123202138 > <...>-409454 [022] .... 28287105.928018: mm_shrink_slab_end: super_cach= e_scan+0x0/0x1a0 ffff9a83046f3458: > nid: 1 unused scan count 3641681686040 new scan count 3641798379189 tot= al_scan 602 > last shrinker return val 123186855 >=20 > The vfs cache and page cache ration was 10:1 on this machine, and half = of caches were dropped. > This also resulted in significant amount of page caches were dropped du= e to inodes eviction. >=20 > Make nr_deferred per memcg for memcg aware shrinkers would solve the un= fairness and bring > better isolation. >=20 > When memcg is not enabled (!CONFIG_MEMCG or memcg disabled), the shrink= er's nr_deferred > would be used. And non memcg aware shrinkers use shrinker's nr_deferre= d all the time. >=20 > Signed-off-by: Yang Shi > --- > include/linux/memcontrol.h | 7 +++--- > mm/vmscan.c | 48 +++++++++++++++++++++++++------------- > 2 files changed, 36 insertions(+), 19 deletions(-) >=20 > diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h > index 62b888b88a5f..e0384367e07d 100644 > --- a/include/linux/memcontrol.h > +++ b/include/linux/memcontrol.h > @@ -93,12 +93,13 @@ struct lruvec_stat { > }; > =20 > /* > - * Bitmap of shrinker::id corresponding to memcg-aware shrinkers, > - * which have elements charged to this memcg. > + * Bitmap and deferred work of shrinker::id corresponding to memcg-awa= re > + * shrinkers, which have elements charged to this memcg. > */ > struct shrinker_info { > struct rcu_head rcu; > - unsigned long map[]; > + unsigned long *map; > + atomic_long_t *nr_deferred; > }; > =20 > /* > diff --git a/mm/vmscan.c b/mm/vmscan.c > index 256896d157d4..20be0db291fe 100644 > --- a/mm/vmscan.c > +++ b/mm/vmscan.c > @@ -187,16 +187,21 @@ static DECLARE_RWSEM(shrinker_rwsem); > #ifdef CONFIG_MEMCG > static int shrinker_nr_max; > =20 > +#define NR_MAX_TO_SHR_MAP_SIZE(nr_max) \ > + ((nr_max / BITS_PER_LONG + 1) * sizeof(unsigned long)) Could have been part of patch 4 already. And yeah, using DIV_ROUND_UP(), = as being hidden in a macro makes the "shorter statement" benefit disappear := ) > + > static void free_shrinker_info_rcu(struct rcu_head *head) > { > kvfree(container_of(head, struct shrinker_info, rcu)); > } > =20 > static int expand_one_shrinker_info(struct mem_cgroup *memcg, > - int size, int old_size) > + int m_size, int d_size, > + int old_m_size, int old_d_size) > { > struct shrinker_info *new, *old; > int nid; > + int size =3D m_size + d_size; > =20 > for_each_node(nid) { > old =3D rcu_dereference_protected( > @@ -209,9 +214,15 @@ static int expand_one_shrinker_info(struct mem_cgr= oup *memcg, > if (!new) > return -ENOMEM; > =20 > - /* Set all old bits, clear all new bits */ > - memset(new->map, (int)0xff, old_size); > - memset((void *)new->map + old_size, 0, size - old_size); > + new->map =3D (unsigned long *)(new + 1); > + new->nr_deferred =3D (void *)new->map + m_size; This better be aligned to sizeof(atomic_long_t). Can we be sure about tha= t? Also it's all quite ugly and complex. Is it worth it? What about just lea= ving map as it is and allocating a nr_deferred array separately, i.e.: struct shrinker_info { struct rcu_head rcu; atomic_long_t *nr_deferred; // allocated separately unsigned long map[]; }; > + /* map: set all old bits, clear all new bits */ > + memset(new->map, (int)0xff, old_m_size); > + memset((void *)new->map + old_m_size, 0, m_size - old_m_size); > + /* nr_deferred: copy old values, clear all new values */ > + memcpy(new->nr_deferred, old->nr_deferred, old_d_size); > + memset((void *)new->nr_deferred + old_d_size, 0, d_size - old_d_size= ); > =20 > rcu_assign_pointer(memcg->nodeinfo[nid]->shrinker_info, new); > call_rcu(&old->rcu, free_shrinker_info_rcu); > @@ -226,9 +237,6 @@ void free_shrinker_info(struct mem_cgroup *memcg) > struct shrinker_info *info; > int nid; > =20 > - if (mem_cgroup_is_root(memcg)) > - return; > - > for_each_node(nid) { > pn =3D mem_cgroup_nodeinfo(memcg, nid); > info =3D rcu_dereference_protected(pn->shrinker_info, true); > @@ -242,12 +250,13 @@ int alloc_shrinker_info(struct mem_cgroup *memcg) > { > struct shrinker_info *info; > int nid, size, ret =3D 0; > - > - if (mem_cgroup_is_root(memcg)) > - return 0; > + int m_size, d_size =3D 0; > =20 > down_write(&shrinker_rwsem); > - size =3D (shrinker_nr_max / BITS_PER_LONG + 1) * sizeof(unsigned long= ); > + m_size =3D NR_MAX_TO_SHR_MAP_SIZE(shrinker_nr_max); > + d_size =3D shrinker_nr_max * sizeof(atomic_long_t); > + size =3D m_size + d_size; > + > for_each_node(nid) { > info =3D kvzalloc_node(sizeof(*info) + size, GFP_KERNEL, nid); > if (!info) { > @@ -255,6 +264,8 @@ int alloc_shrinker_info(struct mem_cgroup *memcg) > ret =3D -ENOMEM; > break; > } > + info->map =3D (unsigned long *)(info + 1); > + info->nr_deferred =3D (void *)info->map + m_size; > rcu_assign_pointer(memcg->nodeinfo[nid]->shrinker_info, info); > } > up_write(&shrinker_rwsem); > @@ -266,10 +277,16 @@ static int expand_shrinker_info(int new_id) > { > int size, old_size, ret =3D 0; > int new_nr_max =3D new_id + 1; > + int m_size, d_size =3D 0; > + int old_m_size, old_d_size =3D 0; > struct mem_cgroup *memcg; > =20 > - size =3D (new_nr_max / BITS_PER_LONG + 1) * sizeof(unsigned long); > - old_size =3D (shrinker_nr_max / BITS_PER_LONG + 1) * sizeof(unsigned = long); > + m_size =3D NR_MAX_TO_SHR_MAP_SIZE(new_nr_max); > + d_size =3D new_nr_max * sizeof(atomic_long_t); > + size =3D m_size + d_size; > + old_m_size =3D NR_MAX_TO_SHR_MAP_SIZE(shrinker_nr_max); > + old_d_size =3D shrinker_nr_max * sizeof(atomic_long_t); > + old_size =3D old_m_size + old_d_size; > if (size <=3D old_size) > goto out; > =20 > @@ -278,9 +295,8 @@ static int expand_shrinker_info(int new_id) > =20 > memcg =3D mem_cgroup_iter(NULL, NULL, NULL); > do { > - if (mem_cgroup_is_root(memcg)) > - continue; > - ret =3D expand_one_shrinker_info(memcg, size, old_size); > + ret =3D expand_one_shrinker_info(memcg, m_size, d_size, > + old_m_size, old_d_size); > if (ret) { > mem_cgroup_iter_break(NULL, memcg); > goto out; >=20