From mboxrd@z Thu Jan 1 00:00:00 1970 From: Roman Gushchin Subject: Re: [PATCH] mm: memcontrol: use per-cpu stocks for socket memory uncharging Date: Wed, 30 Aug 2017 11:55:24 +0100 Message-ID: <20170830105524.GA2852@castle.dhcp.TheFacebook.com> References: <20170829100150.4580-1-guro@fb.com> <20170829192621.GA5447@cmpxchg.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Return-path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=fb.com; h=date : from : to : cc : subject : message-id : references : mime-version : content-type : in-reply-to; s=facebook; bh=NBkUNY16uNSUXEyC00g5pjnAYKnpF+kOqCPXA4dTBU4=; b=jSDpf8l9/sEibonfkrsntiqWlFEcH4+W40rqj9NpmVZQ7nh3dsq618Q805OZlIgfXMvp Yhbfe4ajrH8i4V0unMu6J96UIkKFI1vfmxRKImlCQC8ZYoPriFndHc5xouTsjd2hCTnR /t5Saop4OzurUIWjlJmVoQXWJKxy8o3HQoI= DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=fb.onmicrosoft.com; s=selector1-fb-com; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version; bh=NBkUNY16uNSUXEyC00g5pjnAYKnpF+kOqCPXA4dTBU4=; b=QGF8YxVBkJx7j1m5w8HnpwVBeZbQY6Iot3WvaXEFGAxXLSzpcGo/uyh1/u9RhH9J+3T1FYvTagMUpy1TsThM32DOALys+sBoOeCTRO3gQEoHf5GmcLzHpWY/tBuexmDVsF/w/RgMYxmyTUT8W0Nwtvf3G/PRLrBDysTkDQeXF38= Content-Disposition: inline In-Reply-To: <20170829192621.GA5447-druUgvl0LCNAfugRpC6u6w@public.gmane.org> Sender: cgroups-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Transfer-Encoding: 7bit To: Johannes Weiner Cc: linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org, Michal Hocko , Vladimir Davydov , cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, kernel-team-b10kYP2dOMg@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org On Tue, Aug 29, 2017 at 03:26:21PM -0400, Johannes Weiner wrote: > On Tue, Aug 29, 2017 at 11:01:50AM +0100, Roman Gushchin wrote: > > We've noticed a quite sensible performance overhead on some hosts > > with significant network traffic when socket memory accounting > > is enabled. > > > > Perf top shows that socket memory uncharging path is hot: > > 2.13% [kernel] [k] page_counter_cancel > > 1.14% [kernel] [k] __sk_mem_reduce_allocated > > 1.14% [kernel] [k] _raw_spin_lock > > 0.87% [kernel] [k] _raw_spin_lock_irqsave > > 0.84% [kernel] [k] tcp_ack > > 0.84% [kernel] [k] ixgbe_poll > > 0.83% < workload > > > 0.82% [kernel] [k] enqueue_entity > > 0.68% [kernel] [k] __fget > > 0.68% [kernel] [k] tcp_delack_timer_handler > > 0.67% [kernel] [k] __schedule > > 0.60% < workload > > > 0.59% [kernel] [k] __inet6_lookup_established > > 0.55% [kernel] [k] __switch_to > > 0.55% [kernel] [k] menu_select > > 0.54% libc-2.20.so [.] __memcpy_avx_unaligned > > > > To address this issue, the existing per-cpu stock infrastructure > > can be used. > > > > refill_stock() can be called from mem_cgroup_uncharge_skmem() > > to move charge to a per-cpu stock instead of calling atomic > > page_counter_uncharge(). > > > > To prevent the uncontrolled growth of per-cpu stocks, > > refill_stock() will explicitly drain the cached charge, > > if the cached value exceeds CHARGE_BATCH. > > > > This allows significantly optimize the load: > > 1.21% [kernel] [k] _raw_spin_lock > > 1.01% [kernel] [k] ixgbe_poll > > 0.92% [kernel] [k] _raw_spin_lock_irqsave > > 0.90% [kernel] [k] enqueue_entity > > 0.86% [kernel] [k] tcp_ack > > 0.85% < workload > > > 0.74% perf-11120.map [.] 0x000000000061bf24 > > 0.73% [kernel] [k] __schedule > > 0.67% [kernel] [k] __fget > > 0.63% [kernel] [k] __inet6_lookup_established > > 0.62% [kernel] [k] menu_select > > 0.59% < workload > > > 0.59% [kernel] [k] __switch_to > > 0.57% libc-2.20.so [.] __memcpy_avx_unaligned > > > > Signed-off-by: Roman Gushchin > > Cc: Johannes Weiner > > Cc: Michal Hocko > > Cc: Vladimir Davydov > > Cc: cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org > > Cc: kernel-team-b10kYP2dOMg@public.gmane.org > > Cc: linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org > > Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org > > Acked-by: Johannes Weiner > > Neat! > > As far as other types of pages go: page cache and anon are already > batched pretty well, but I think kmem might benefit from this > too. Have you considered using the stock in memcg_kmem_uncharge()? Good idea! I'll try to find an appropriate testcase and check if it really brings any benefits. If so, I'll master a patch. Thanks!