From: Michal Hocko <mhocko@suse.com> To: Roman Gushchin <roman.gushchin@linux.dev> Cc: "Marcelo Tosatti" <mtosatti@redhat.com>, "Leonardo Brás" <leobras@redhat.com>, "Johannes Weiner" <hannes@cmpxchg.org>, "Shakeel Butt" <shakeelb@google.com>, "Muchun Song" <muchun.song@linux.dev>, "Andrew Morton" <akpm@linux-foundation.org>, cgroups@vger.kernel.org, linux-mm@kvack.org, linux-kernel@vger.kernel.org, "Frederic Weisbecker" <fweisbecker@suse.de> Subject: Re: [PATCH v2 0/5] Introduce memcg_stock_pcp remote draining Date: Fri, 27 Jan 2023 08:11:04 +0100 [thread overview] Message-ID: <Y9N5CI8PpsfiaY9c@dhcp22.suse.cz> (raw) In-Reply-To: <Y9MI42NSLooyVZNu@P9FQF9L96D.corp.robot.car> [Cc Frederic] On Thu 26-01-23 15:12:35, Roman Gushchin wrote: > On Thu, Jan 26, 2023 at 08:41:34AM +0100, Michal Hocko wrote: [...] > > > Essentially each cpu will try to grab the remains of the memory quota > > > and move it locally. I wonder in such circumstances if we need to disable the pcp-caching > > > on per-cgroup basis. > > > > I think it would be more than sufficient to disable pcp charging on an > > isolated cpu. > > It might have significant performance consequences. Is it really significant? > I'd rather opt out of stock draining for isolated cpus: it might slightly reduce > the accuracy of memory limits and slightly increase the memory footprint (all > those dying memcgs...), but the impact will be limited. Actually it is limited > by the number of cpus. Hmm, OK, I have misunderstood your proposal. Yes, the overal pcp charges potentially left behind should be small and that shouldn't really be a concern for memcg oom situations (unless the limit is very small and workloads on isolated cpus using small hard limits is way beyond my imagination). My first thought was that those charges could be left behind without any upper bound but in reality sooner or later something should be running on those cpus and if the memcg is gone the pcp cache would get refilled and old charges gone. So yes, this is actually a better and even simpler solution. All we need is something like this diff --git a/mm/memcontrol.c b/mm/memcontrol.c index ab457f0394ab..13b84bbd70ba 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -2344,6 +2344,9 @@ static void drain_all_stock(struct mem_cgroup *root_memcg) struct mem_cgroup *memcg; bool flush = false; + if (cpu_is_isolated(cpu)) + continue; + rcu_read_lock(); memcg = stock->cached; if (memcg && stock->nr_pages && There is no such cpu_is_isolated() AFAICS so we would need a help from NOHZ and cpuisol people to create one for us. Frederic, would such an abstraction make any sense from your POV? -- Michal Hocko SUSE Labs
WARNING: multiple messages have this Message-ID (diff)
From: Michal Hocko <mhocko-IBi9RG/b67k@public.gmane.org> To: Roman Gushchin <roman.gushchin-fxUVXftIFDnyG1zEObXtfA@public.gmane.org> Cc: "Marcelo Tosatti" <mtosatti-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>, "Leonardo Brás" <leobras-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>, "Johannes Weiner" <hannes-druUgvl0LCNAfugRpC6u6w@public.gmane.org>, "Shakeel Butt" <shakeelb-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org>, "Muchun Song" <muchun.song-fxUVXftIFDnyG1zEObXtfA@public.gmane.org>, "Andrew Morton" <akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>, cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, "Frederic Weisbecker" <fweisbecker-l3A5Bk7waGM@public.gmane.org> Subject: Re: [PATCH v2 0/5] Introduce memcg_stock_pcp remote draining Date: Fri, 27 Jan 2023 08:11:04 +0100 [thread overview] Message-ID: <Y9N5CI8PpsfiaY9c@dhcp22.suse.cz> (raw) In-Reply-To: <Y9MI42NSLooyVZNu-+xijCwNIfdoLQcUKs7qKB+WAnPUfkyWGUBSOeVevoDU@public.gmane.org> [Cc Frederic] On Thu 26-01-23 15:12:35, Roman Gushchin wrote: > On Thu, Jan 26, 2023 at 08:41:34AM +0100, Michal Hocko wrote: [...] > > > Essentially each cpu will try to grab the remains of the memory quota > > > and move it locally. I wonder in such circumstances if we need to disable the pcp-caching > > > on per-cgroup basis. > > > > I think it would be more than sufficient to disable pcp charging on an > > isolated cpu. > > It might have significant performance consequences. Is it really significant? > I'd rather opt out of stock draining for isolated cpus: it might slightly reduce > the accuracy of memory limits and slightly increase the memory footprint (all > those dying memcgs...), but the impact will be limited. Actually it is limited > by the number of cpus. Hmm, OK, I have misunderstood your proposal. Yes, the overal pcp charges potentially left behind should be small and that shouldn't really be a concern for memcg oom situations (unless the limit is very small and workloads on isolated cpus using small hard limits is way beyond my imagination). My first thought was that those charges could be left behind without any upper bound but in reality sooner or later something should be running on those cpus and if the memcg is gone the pcp cache would get refilled and old charges gone. So yes, this is actually a better and even simpler solution. All we need is something like this diff --git a/mm/memcontrol.c b/mm/memcontrol.c index ab457f0394ab..13b84bbd70ba 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -2344,6 +2344,9 @@ static void drain_all_stock(struct mem_cgroup *root_memcg) struct mem_cgroup *memcg; bool flush = false; + if (cpu_is_isolated(cpu)) + continue; + rcu_read_lock(); memcg = stock->cached; if (memcg && stock->nr_pages && There is no such cpu_is_isolated() AFAICS so we would need a help from NOHZ and cpuisol people to create one for us. Frederic, would such an abstraction make any sense from your POV? -- Michal Hocko SUSE Labs
next prev parent reply other threads:[~2023-01-27 7:11 UTC|newest] Thread overview: 85+ messages / expand[flat|nested] mbox.gz Atom feed top 2023-01-25 7:34 [PATCH v2 0/5] Introduce memcg_stock_pcp remote draining Leonardo Bras 2023-01-25 7:34 ` Leonardo Bras 2023-01-25 7:34 ` [PATCH v2 1/5] mm/memcontrol: Align percpu memcg_stock to cache Leonardo Bras 2023-01-25 7:34 ` Leonardo Bras 2023-01-25 7:34 ` [PATCH v2 2/5] mm/memcontrol: Change stock_lock type from local_lock_t to spinlock_t Leonardo Bras 2023-01-25 7:34 ` Leonardo Bras 2023-01-25 7:35 ` [PATCH v2 3/5] mm/memcontrol: Reorder memcg_stock_pcp members to avoid holes Leonardo Bras 2023-01-25 7:35 ` Leonardo Bras 2023-01-25 7:35 ` [PATCH v2 4/5] mm/memcontrol: Perform all stock drain in current CPU Leonardo Bras 2023-01-25 7:35 ` [PATCH v2 5/5] mm/memcontrol: Remove flags from memcg_stock_pcp Leonardo Bras 2023-01-25 7:35 ` Leonardo Bras 2023-01-25 8:33 ` [PATCH v2 0/5] Introduce memcg_stock_pcp remote draining Michal Hocko 2023-01-25 8:33 ` Michal Hocko 2023-01-25 11:06 ` Leonardo Brás 2023-01-25 11:06 ` Leonardo Brás 2023-01-25 11:39 ` Michal Hocko 2023-01-25 11:39 ` Michal Hocko 2023-01-25 18:22 ` Marcelo Tosatti 2023-01-25 18:22 ` Marcelo Tosatti 2023-01-25 23:14 ` Roman Gushchin 2023-01-25 23:14 ` Roman Gushchin 2023-01-26 7:41 ` Michal Hocko 2023-01-26 7:41 ` Michal Hocko 2023-01-26 18:03 ` Marcelo Tosatti 2023-01-26 19:20 ` Michal Hocko 2023-01-27 0:32 ` Marcelo Tosatti 2023-01-27 0:32 ` Marcelo Tosatti 2023-01-27 6:58 ` Michal Hocko 2023-01-27 6:58 ` Michal Hocko 2023-02-01 18:31 ` Roman Gushchin 2023-01-26 23:12 ` Roman Gushchin 2023-01-27 7:11 ` Michal Hocko [this message] 2023-01-27 7:11 ` Michal Hocko 2023-01-27 7:22 ` Leonardo Brás 2023-01-27 8:12 ` Leonardo Brás 2023-01-27 8:12 ` Leonardo Brás 2023-01-27 9:23 ` Michal Hocko 2023-01-27 9:23 ` Michal Hocko 2023-01-27 13:03 ` Frederic Weisbecker 2023-01-27 13:03 ` Frederic Weisbecker 2023-01-27 13:58 ` Michal Hocko 2023-01-27 13:58 ` Michal Hocko 2023-01-27 18:18 ` Roman Gushchin 2023-01-27 18:18 ` Roman Gushchin 2023-02-03 15:21 ` Michal Hocko 2023-02-03 15:21 ` Michal Hocko 2023-02-03 19:25 ` Roman Gushchin 2023-02-03 19:25 ` Roman Gushchin 2023-02-13 13:36 ` Michal Hocko 2023-02-13 13:36 ` Michal Hocko 2023-01-27 7:14 ` Leonardo Brás 2023-01-27 7:14 ` Leonardo Brás 2023-01-27 7:20 ` Michal Hocko 2023-01-27 7:20 ` Michal Hocko 2023-01-27 7:35 ` Leonardo Brás 2023-01-27 9:29 ` Michal Hocko 2023-01-27 19:29 ` Leonardo Brás 2023-01-27 19:29 ` Leonardo Brás 2023-01-27 23:50 ` Roman Gushchin 2023-01-26 18:19 ` Marcelo Tosatti 2023-01-26 18:19 ` Marcelo Tosatti 2023-01-27 5:40 ` Leonardo Brás 2023-01-27 5:40 ` Leonardo Brás 2023-01-26 2:01 ` Hillf Danton 2023-01-26 7:45 ` Michal Hocko 2023-01-26 7:45 ` Michal Hocko 2023-01-26 18:14 ` Marcelo Tosatti 2023-01-26 18:14 ` Marcelo Tosatti 2023-01-26 19:13 ` Michal Hocko 2023-01-26 19:13 ` Michal Hocko 2023-01-27 6:55 ` Leonardo Brás 2023-01-27 6:55 ` Leonardo Brás 2023-01-31 11:35 ` Marcelo Tosatti 2023-01-31 11:35 ` Marcelo Tosatti 2023-02-01 4:36 ` Leonardo Brás 2023-02-01 12:52 ` Michal Hocko 2023-02-01 12:52 ` Michal Hocko 2023-02-01 12:41 ` Michal Hocko 2023-02-01 12:41 ` Michal Hocko 2023-02-04 4:55 ` Leonardo Brás 2023-02-04 4:55 ` Leonardo Brás 2023-02-05 19:49 ` Roman Gushchin 2023-02-07 3:18 ` Leonardo Brás 2023-02-08 19:23 ` Roman Gushchin 2023-02-08 19:23 ` Roman Gushchin
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=Y9N5CI8PpsfiaY9c@dhcp22.suse.cz \ --to=mhocko@suse.com \ --cc=akpm@linux-foundation.org \ --cc=cgroups@vger.kernel.org \ --cc=fweisbecker@suse.de \ --cc=hannes@cmpxchg.org \ --cc=leobras@redhat.com \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-mm@kvack.org \ --cc=mtosatti@redhat.com \ --cc=muchun.song@linux.dev \ --cc=roman.gushchin@linux.dev \ --cc=shakeelb@google.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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.