From patchwork Tue Mar 28 06:16:29 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Yosry Ahmed X-Patchwork-Id: 13190543 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 64E15C76195 for ; Tue, 28 Mar 2023 06:17:03 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232440AbjC1GRB (ORCPT ); Tue, 28 Mar 2023 02:17:01 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:34378 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232140AbjC1GRA (ORCPT ); Tue, 28 Mar 2023 02:17:00 -0400 Received: from mail-yw1-x114a.google.com (mail-yw1-x114a.google.com [IPv6:2607:f8b0:4864:20::114a]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id B79BB2D76 for ; Mon, 27 Mar 2023 23:16:42 -0700 (PDT) Received: by mail-yw1-x114a.google.com with SMTP id 00721157ae682-541942bfdccso110026227b3.14 for ; Mon, 27 Mar 2023 23:16:42 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20210112; t=1679984202; h=cc:to:from:subject:message-id:mime-version:date:from:to:cc:subject :date:message-id:reply-to; bh=sp9O7DL9obBb0muMtifinlg/m6vcys5sIiJC9lNDWi4=; b=NeuDT5i+IX7EsYuGxrlQHGfM/O44K/q8azh0N01/JWHajTpr/8bCoJ2QTmweAtvhfe gYoeDE3hEA/7en26XdMklOvz+naw4YK/G+YhpbjekM9lK1gn62mOub1rqZqAuOD3Va4S 1QIsws0AsakJv6xmZVg2tR6CGN3YqUveLDQmgizdGWo93cuUB1v+gh6AulTa+cAtsQ1E KKArWRQY0GrdStBwfUXdbR6yPsdQlbb61EZd9PCwqGtcxA7k1j294s/M/nt3EKFywNtt l3OyHVKVYr6ScGHLswWLNtnzcPUbA+BqIwaG4p3phWhy3WEkn1PyyK7m1rhPq1yMmAEF q2mw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; t=1679984202; h=cc:to:from:subject:message-id:mime-version:date:x-gm-message-state :from:to:cc:subject:date:message-id:reply-to; bh=sp9O7DL9obBb0muMtifinlg/m6vcys5sIiJC9lNDWi4=; b=7iZXqUlQg1tsB8EBVMdhsv29yDGBAf2gIbn44pONpiDIrrJTTDZBW/cEIT2UaVuRXX HT5Yb8j7obJomexJnG0x2rXGd4jhRj/OKE92ZeyqIXepwXHB+IDj7IHvkHb9dkhmdwh7 oMTAk1RBywyH7IICZwvbAjYKz08HvlW87vVzAXgaWAOAc0jE50m0+KEKzf7oOYZDDZNX zgJOobUcP4pgLuV3DcN5f/nCUY/KTAOxEXIn9Llw6EvsvPOPTsWq95JsMbaqoE0+Yvww f4P76W5SGW2tXhbdx5HEiZi7B2T8g1j+DfYFzRXJuqfLy3ULRFQFHm9vm1MlkrceuWJI K+Cg== X-Gm-Message-State: AAQBX9cQ0Q3DVj8tugha4mDeodxgy8BAFaK5g0h/m1DXmEgm0guVvjOv UT0f6eiw7t+Bvw9m0YPtbeZ7OzU0/+TNix+y X-Google-Smtp-Source: AKy350bZY7FA4NsJ0fqBTDQ1vhS8yI9N8IskrUxgaH2uFCVGGGkuCT76RaKDFHndt2HqiHyVKmyLU1Kd/HHTyvK/ X-Received: from yosry.c.googlers.com ([fda3:e722:ac3:cc00:7f:e700:c0a8:2327]) (user=yosryahmed job=sendgmr) by 2002:a25:2749:0:b0:b4a:e062:3576 with SMTP id n70-20020a252749000000b00b4ae0623576mr6830636ybn.13.1679984201935; Mon, 27 Mar 2023 23:16:41 -0700 (PDT) Date: Tue, 28 Mar 2023 06:16:29 +0000 Mime-Version: 1.0 X-Mailer: git-send-email 2.40.0.348.gf938b09366-goog Message-ID: <20230328061638.203420-1-yosryahmed@google.com> Subject: [PATCH v1 0/9] memcg: make rstat flushing irq and sleep friendly From: Yosry Ahmed To: Tejun Heo , Josef Bacik , Jens Axboe , Zefan Li , Johannes Weiner , Michal Hocko , Roman Gushchin , Shakeel Butt , Muchun Song , Andrew Morton , " =?utf-8?q?Michal_Koutn=C3=BD?= " Cc: Vasily Averin , cgroups@vger.kernel.org, linux-block@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, bpf@vger.kernel.org, Yosry Ahmed Precedence: bulk List-ID: X-Mailing-List: bpf@vger.kernel.org Currently, all calls to flush memcg stats use the atomic variant for rstat flushing, cgroup_rstat_flush_irqsafe(), which keeps interrupts disabled throughout flushing and does not sleep. Flushing stats is an expensive operation, and we should avoid doing it atomically where possible. Otherwise, we may end up doing a lot of work without rescheduling and with interrupts disabled unnecessarily. Patches 1 and 2 are cleanups requested during reviews of prior versions of this series. Patch 3 makes sure we never try to flush from within an irq context, and patch 4 adds a WARN_ON_ONCE() to make sure we catch any violations. Patches 5 to 8 introduce separate variants of mem_cgroup_flush_stats() for atomic and non-atomic flushing, and make sure we only flush the stats atomically when necessary. Patch 9 is a slightly tangential optimization that limits the work done by rstat flushing in some scenarios. RFC -> v1: - Dropped patch 1 that attempted to make the global rstat lock a non-irq lock, will follow up on that separetly (Shakeel). - Dropped stats_flush_lock entirely, replaced by an atomic (Johannes). - Renamed cgroup_rstat_flush_irqsafe() to cgroup_rstat_flush_atomic() instead of removing it (Johannes). - Added a patch to rename mem_cgroup_flush_stats_delayed() to mem_cgroup_flush_stats_ratelimited() (Johannes). - Separate APIs for flushing memcg stats in atomic and non-atomic contexts instead of a boolean argument (Johannes). - Added patches 3 & 4 to make sure we never flush from irq context (Shakeel & Johannes). Yosry Ahmed (9): cgroup: rename cgroup_rstat_flush_"irqsafe" to "atomic" memcg: rename mem_cgroup_flush_stats_"delayed" to "ratelimited" memcg: do not flush stats in irq context cgroup: rstat: add WARN_ON_ONCE() if flushing outside task context memcg: replace stats_flush_lock with an atomic memcg: sleep during flushing stats in safe contexts workingset: memcg: sleep when flushing stats in workingset_refault() vmscan: memcg: sleep when flushing stats during reclaim memcg: do not modify rstat tree for zero updates include/linux/cgroup.h | 2 +- include/linux/memcontrol.h | 9 +++- kernel/cgroup/rstat.c | 6 ++- mm/memcontrol.c | 86 ++++++++++++++++++++++++++++++++------ mm/workingset.c | 4 +- 5 files changed, 87 insertions(+), 20 deletions(-)