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 Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 3FC9BC43217 for ; Tue, 8 Feb 2022 08:19:55 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1349044AbiBHITx (ORCPT ); Tue, 8 Feb 2022 03:19:53 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:53232 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1349158AbiBHITr (ORCPT ); Tue, 8 Feb 2022 03:19:47 -0500 Received: from mail-yb1-xb49.google.com (mail-yb1-xb49.google.com [IPv6:2607:f8b0:4864:20::b49]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 65899C03FED4 for ; Tue, 8 Feb 2022 00:19:39 -0800 (PST) Received: by mail-yb1-xb49.google.com with SMTP id f32-20020a25b0a0000000b0061dad37dcd6so11423221ybj.16 for ; Tue, 08 Feb 2022 00:19:39 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20210112; h=date:in-reply-to:message-id:mime-version:references:subject:from:to :cc:content-transfer-encoding; bh=lRNGv978IZUzuZtn7+GJfB1JhMWHf1r6lSn8z5WI1mY=; b=INp8EKCPNwNrzFQ2lW8+ZqMYhAtdAEXCXLlJVllljVDNsoGTRm2ZuYxrtOcHVsp/hJ qmssG44HpUq64mP3uLn/4AVI1ej7zkoKRZWpxXkOxhxavN0fQ+I9m7cy7g8P97wLPGBw NY0xshp2tLyOPGdrOUuB5HoeJdMea7eAJE8C4wvmp6nNQciiFuE+WjSFzsZM5l3dFI8A 4dQU9Ou2fEJQqJbBAdBnKNy6nOdDXSgR7uVtCvlcRb795L8OEurajARCMGpgCZxjqeQP sEetj61pS6mz0Xca5MGDgAbECJWqr+tKRTa9ZXEmvfYTizKbzD38q8OtNntJi9hD/hoC tL4w== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:date:in-reply-to:message-id:mime-version :references:subject:from:to:cc:content-transfer-encoding; bh=lRNGv978IZUzuZtn7+GJfB1JhMWHf1r6lSn8z5WI1mY=; b=XWTmel0iTs8snmutffCzr/Sdd7zIRQ5AlIGBVpfsah8eV+I7W/U4yeONbouBWLVh22 lhpu8v2YSN3m3jfO7cMR3TqNFTh+n8qbcj54YJenKbwhc3QNoyLaHUpTPy08O1RC/PZW r7FMwyFIrq5K5ZUIEmHnGh9ErtY9IqUe+qn7i7hLpGnCR0l4WetXCvSKyiCTJliZi/PN M+/g0o6kUKVxtAoHY0k1A5hDZhCDqk2BypQV5EALF8Vj+qBGy9hpLGPYxhNP7llv83ej 8H4I43rIumHRvkvYbHxWMWXEY9Bf9h11a0mVfccLm5SYPEn1Q8hTE0DD//JzlZM66etw dZ2g== X-Gm-Message-State: AOAM531XiUQYDvbVx6Mwv0UIDACc0d2g/39ZhfmUKmVrKm42oso7f6j2 YCkzMl9gwPLMvXr1BzBzcDBVp+UFFSw= X-Google-Smtp-Source: ABdhPJxJg59mt99jNZX53wlCHH+5BfJiYUEn9IVqtde0TbSBN1Ev7xHE2KNmWqJImN86wGIDNgk+93Hy6EU= X-Received: from yuzhao.bld.corp.google.com ([2620:15c:183:200:5f31:19c3:21f5:7300]) (user=yuzhao job=sendgmr) by 2002:a81:3346:: with SMTP id z67mr3736711ywz.187.1644308378463; Tue, 08 Feb 2022 00:19:38 -0800 (PST) Date: Tue, 8 Feb 2022 01:18:58 -0700 In-Reply-To: <20220208081902.3550911-1-yuzhao@google.com> Message-Id: <20220208081902.3550911-9-yuzhao@google.com> Mime-Version: 1.0 References: <20220208081902.3550911-1-yuzhao@google.com> X-Mailer: git-send-email 2.35.0.263.gb82422642f-goog Subject: [PATCH v7 08/12] mm: multigenerational LRU: optimize multiple memcgs From: Yu Zhao To: Andrew Morton , Johannes Weiner , Mel Gorman , Michal Hocko Cc: Andi Kleen , Aneesh Kumar , Barry Song <21cnbao@gmail.com>, Catalin Marinas , Dave Hansen , Hillf Danton , Jens Axboe , Jesse Barnes , Jonathan Corbet , Linus Torvalds , Matthew Wilcox , Michael Larabel , Mike Rapoport , Rik van Riel , Vlastimil Babka , Will Deacon , Ying Huang , linux-arm-kernel@lists.infradead.org, linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, page-reclaim@google.com, x86@kernel.org, Yu Zhao , Brian Geffon , Jan Alexander Steffens , Oleksandr Natalenko , Steven Barrett , Suleiman Souhlal , Daniel Byrne , Donald Carr , "=?UTF-8?q?Holger=20Hoffst=C3=A4tte?=" , Konstantin Kharlamov , Shuang Zhai , Sofia Trinh Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-doc@vger.kernel.org When multiple memcgs are available, it's possible to improve the overall performance under global memory pressure by making better choices based on generations and tiers. This patch adds a rudimentary optimization to select memcgs that can drop single-use unmapped clean pages first, and thus it reduces the chance of going into the aging path or swapping, which can be costly. Its goal is to improve the overall performance when there are mixed types of workloads, e.g., heavy anon workload in one memcg and heavy buffered I/O workload in the other. Though this optimization can be applied to both kswapd and direct reclaim, it's only added to kswapd to keep the patchset manageable. Later improvements will cover the direct reclaim path. Server benchmark results: Mixed workloads: fio (buffered I/O): -[28, 30]% IOPS BW patch1-7: 3117k 11.9GiB/s patch1-8: 2217k 8661MiB/s memcached (anon): +[247, 251]% Ops/sec KB/sec patch1-7: 563772.35 21900.01 patch1-8: 1968343.76 76461.24 Mixed workloads: fio (buffered I/O): -[4, 6]% IOPS BW 5.17-rc2: 2338k 9133MiB/s patch1-8: 2217k 8661MiB/s memcached (anon): +[524, 530]% Ops/sec KB/sec 5.17-rc2: 313821.65 12190.55 patch1-8: 1968343.76 76461.24 Configurations: (changes since patch 5) cat combined.sh modprobe brd rd_nr=3D2 rd_size=3D56623104 swapoff -a mkswap /dev/ram0 swapon /dev/ram0 mkfs.ext4 /dev/ram1 mount -t ext4 /dev/ram1 /mnt memtier_benchmark -S /var/run/memcached/memcached.sock \ -P memcache_binary -n allkeys --key-minimum=3D1 \ --key-maximum=3D50000000 --key-pattern=3DP:P -c 1 -t 36 \ --ratio 1:0 --pipeline 8 -d 2000 fio -name=3Dmglru --numjobs=3D36 --directory=3D/mnt --size=3D1408m \ --buffered=3D1 --ioengine=3Dio_uring --iodepth=3D128 \ --iodepth_batch_submit=3D32 --iodepth_batch_complete=3D32 \ --rw=3Drandread --random_distribution=3Drandom --norandommap \ --time_based --ramp_time=3D10m --runtime=3D90m --group_reporting & pid=3D$! sleep 200 memtier_benchmark -S /var/run/memcached/memcached.sock \ -P memcache_binary -n allkeys --key-minimum=3D1 \ --key-maximum=3D50000000 --key-pattern=3DR:R -c 1 -t 36 \ --ratio 0:1 --pipeline 8 --randomize --distinct-client-seed kill -INT $pid wait Client benchmark results: no change (CONFIG_MEMCG=3Dn) Signed-off-by: Yu Zhao Acked-by: Brian Geffon Acked-by: Jan Alexander Steffens (heftig) Acked-by: Oleksandr Natalenko Acked-by: Steven Barrett Acked-by: Suleiman Souhlal Tested-by: Daniel Byrne Tested-by: Donald Carr Tested-by: Holger Hoffst=C3=A4tte Tested-by: Konstantin Kharlamov Tested-by: Shuang Zhai Tested-by: Sofia Trinh --- mm/vmscan.c | 45 +++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 41 insertions(+), 4 deletions(-) diff --git a/mm/vmscan.c b/mm/vmscan.c index 5ab6cd332fcc..fc09b6c10624 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -127,6 +127,13 @@ struct scan_control { /* Always discard instead of demoting to lower tier memory */ unsigned int no_demotion:1; =20 +#ifdef CONFIG_LRU_GEN + /* help make better choices when multiple memcgs are available */ + unsigned int memcgs_need_aging:1; + unsigned int memcgs_need_swapping:1; + unsigned int memcgs_avoid_swapping:1; +#endif + /* Allocation order */ s8 order; =20 @@ -4343,6 +4350,22 @@ static void lru_gen_age_node(struct pglist_data *pgd= at, struct scan_control *sc) =20 VM_BUG_ON(!current_is_kswapd()); =20 + /* + * To reduce the chance of going into the aging path or swapping, which + * can be costly, optimistically skip them unless their corresponding + * flags were cleared in the eviction path. This improves the overall + * performance when multiple memcgs are available. + */ + if (!sc->memcgs_need_aging) { + sc->memcgs_need_aging =3D true; + sc->memcgs_avoid_swapping =3D !sc->memcgs_need_swapping; + sc->memcgs_need_swapping =3D true; + return; + } + + sc->memcgs_need_swapping =3D true; + sc->memcgs_avoid_swapping =3D true; + current->reclaim_state->mm_walk =3D &pgdat->mm_walk; =20 memcg =3D mem_cgroup_iter(NULL, NULL, NULL); @@ -4745,7 +4768,8 @@ static int isolate_folios(struct lruvec *lruvec, stru= ct scan_control *sc, int sw return scanned; } =20 -static int evict_folios(struct lruvec *lruvec, struct scan_control *sc, in= t swappiness) +static int evict_folios(struct lruvec *lruvec, struct scan_control *sc, in= t swappiness, + bool *swapped) { int type; int scanned; @@ -4810,6 +4834,9 @@ static int evict_folios(struct lruvec *lruvec, struct= scan_control *sc, int swap =20 sc->nr_reclaimed +=3D reclaimed; =20 + if (!type && swapped) + *swapped =3D true; + return scanned; } =20 @@ -4838,8 +4865,10 @@ static long get_nr_to_scan(struct lruvec *lruvec, st= ruct scan_control *sc, bool if (!nr_to_scan) return 0; =20 - if (!need_aging) + if (!need_aging) { + sc->memcgs_need_aging =3D false; return nr_to_scan; + } =20 /* leave the work to lru_gen_age_node() */ if (current_is_kswapd()) @@ -4861,6 +4890,8 @@ static void lru_gen_shrink_lruvec(struct lruvec *lruv= ec, struct scan_control *sc { struct blk_plug plug; long scanned =3D 0; + bool swapped =3D false; + unsigned long reclaimed =3D sc->nr_reclaimed; struct mem_cgroup *memcg =3D lruvec_memcg(lruvec); struct pglist_data *pgdat =3D lruvec_pgdat(lruvec); =20 @@ -4887,13 +4918,19 @@ static void lru_gen_shrink_lruvec(struct lruvec *lr= uvec, struct scan_control *sc if (!nr_to_scan) break; =20 - delta =3D evict_folios(lruvec, sc, swappiness); + delta =3D evict_folios(lruvec, sc, swappiness, &swapped); if (!delta) break; =20 + if (sc->memcgs_avoid_swapping && swappiness < 200 && swapped) + break; + scanned +=3D delta; - if (scanned >=3D nr_to_scan) + if (scanned >=3D nr_to_scan) { + if (!swapped && sc->nr_reclaimed - reclaimed >=3D MIN_LRU_BATCH) + sc->memcgs_need_swapping =3D false; break; + } =20 cond_resched(); } --=20 2.35.0.263.gb82422642f-goog