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 kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by smtp.lore.kernel.org (Postfix) with ESMTP id 43D8ACA0FE1 for ; Fri, 1 Sep 2023 06:24:11 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id C5F898D001E; Fri, 1 Sep 2023 02:24:10 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id C36AC8D0002; Fri, 1 Sep 2023 02:24:10 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id B30968D001E; Fri, 1 Sep 2023 02:24:10 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0016.hostedemail.com [216.40.44.16]) by kanga.kvack.org (Postfix) with ESMTP id A2EC28D0002 for ; Fri, 1 Sep 2023 02:24:10 -0400 (EDT) Received: from smtpin18.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay03.hostedemail.com (Postfix) with ESMTP id 7780CA0146 for ; Fri, 1 Sep 2023 06:24:10 +0000 (UTC) X-FDA: 81187038660.18.802F015 Received: from mail-pf1-f181.google.com (mail-pf1-f181.google.com [209.85.210.181]) by imf29.hostedemail.com (Postfix) with ESMTP id AD460120012 for ; Fri, 1 Sep 2023 06:24:08 +0000 (UTC) Authentication-Results: imf29.hostedemail.com; dkim=pass header.d=bytedance.com header.s=google header.b="LoAjBs/W"; dmarc=pass (policy=quarantine) header.from=bytedance.com; spf=pass (imf29.hostedemail.com: domain of wuyun.abel@bytedance.com designates 209.85.210.181 as permitted sender) smtp.mailfrom=wuyun.abel@bytedance.com ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1693549448; a=rsa-sha256; cv=none; b=PBmiJFqN4V8QjJwO9Z3mNutV5jgObQuB/jlICEmFpJM5hl6r45Z77UctT9pL0qoezAAdLb 3smyY9ICKwkMdQs3QUnWzaOyPbA8BuXZlVeCEumbKbwwrXTFRgnm8P7llkKPZJBHBTC8o6 WShyXGPTvQfXRY0qzIeqEnM9kA61UC8= ARC-Authentication-Results: i=1; imf29.hostedemail.com; dkim=pass header.d=bytedance.com header.s=google header.b="LoAjBs/W"; dmarc=pass (policy=quarantine) header.from=bytedance.com; spf=pass (imf29.hostedemail.com: domain of wuyun.abel@bytedance.com designates 209.85.210.181 as permitted sender) smtp.mailfrom=wuyun.abel@bytedance.com ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1693549448; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:dkim-signature; bh=ZQ6PDv993Rp2fl4j4E1LU63DQukXqNgX8LZTNx5xQ8Y=; b=GXbc3ihB1QkkJNJaLvJpsRiK0UjkB5FdAngKPaxIOYPtsIec6i6xCI2yUJQa9zH6w51H/q uo47Wu8D5i53Y1VOhq8KRWCIzWWgSyITRN9sbgM2vVjttrtF8V+PMs1g/o3UmXWlAOYvJ/ Rh3t/XeD4hafyRhjgVoGGFsDlCVXV50= Received: by mail-pf1-f181.google.com with SMTP id d2e1a72fcca58-68a3b66f350so1389129b3a.3 for ; Thu, 31 Aug 2023 23:24:08 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=bytedance.com; s=google; t=1693549447; x=1694154247; darn=kvack.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=ZQ6PDv993Rp2fl4j4E1LU63DQukXqNgX8LZTNx5xQ8Y=; b=LoAjBs/W/JDRV2+RlrQwMhDK4UxyIAPcsckN8Br2JY123rqDVzHVgwxSMlKCCT31/s SOiO3Xkf1joTQVBCUkisn/Uj1kX2pUt5/DMK4Gn1hmB43rh0MulR5Prxcaq6bQ8xjDfP 2b9/82VDaWPrDJDqLHOilbpqgcrCP8XEe6zIhWzC/0dCrK0topbSzoWjsLaRiLbPqoO+ yNHcNy4tFxJzivBvobSu3m5hUjlZMUb2bqJ+ZSETzIEC3PsOLSGvATI37LHvJpY4kNcw XkUVXNkpuEPeWvDDJ0laYlekwQr7hPyBdnalcDXWRK+BNgQWBof4YjUhU+mD1fRg75gI IP5Q== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1693549447; x=1694154247; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=ZQ6PDv993Rp2fl4j4E1LU63DQukXqNgX8LZTNx5xQ8Y=; b=DULe+iMqDqEda4AWLcWXJcGoc9J7vHpGdaZywnEH7FQ2IDrMngFg+1frFbpoPqffw8 uRQLMcOtuy9F8+D9IiMqBa+Im61LnqbwJQE0RvEA4QyPGZiHMeK1ryxuR5+Ncc8aKYE6 lidh5klRFbLoznr20d8jKDbXUnygPDpWfsi2g5/iZI4ooL/Z1zdvkMA79fUrpndiaOl0 l8xDkJjTgJA657JOk+gibsIXtOY86JPNr8ZoC4NGd5SUByD1fmTNRlBy0ExAbsD7Q4BO naMoTfLah5S+3FABKobrPtguHj5SOx78zudX9Hy4koX5oGg39Lm0cNrwTWgKK4NNP91V /a4Q== X-Gm-Message-State: AOJu0YwlieKyzNju1kZS0me1dbQOvm4r0NtMrshrgvYS+yPxrN27CIu6 UUnITFLx7Yclvp6x5epTTz9YHg== X-Google-Smtp-Source: AGHT+IFMqQT61UT4Tg0IU3qJAIzxQXgO4DhEf5zuPhELzVvTcr8/BRgzyzW20KePzBL0PUk/x6gtKA== X-Received: by 2002:a05:6a00:17a8:b0:68a:66b2:3d15 with SMTP id s40-20020a056a0017a800b0068a66b23d15mr2442799pfg.7.1693549447179; Thu, 31 Aug 2023 23:24:07 -0700 (PDT) Received: from C02DV8HUMD6R.bytedance.net ([203.208.167.146]) by smtp.gmail.com with ESMTPSA id fm19-20020a056a002f9300b0068c1ac1784csm2223265pfb.59.2023.08.31.23.23.59 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Thu, 31 Aug 2023 23:24:06 -0700 (PDT) From: Abel Wu To: "David S. Miller" , Eric Dumazet , Jakub Kicinski , Paolo Abeni , Andrew Morton , Shakeel Butt , Roman Gushchin , Michal Hocko , Johannes Weiner , Yosry Ahmed , "Matthew Wilcox (Oracle)" , Yu Zhao , Abel Wu , Yafang Shao , Kefeng Wang , Kuniyuki Iwashima , Martin KaFai Lau , Breno Leitao , Alexander Mikhalitsyn , David Howells , Jason Xing Cc: linux-kernel@vger.kernel.org (open list), netdev@vger.kernel.org (open list:NETWORKING [GENERAL]), linux-mm@kvack.org (open list:MEMORY MANAGEMENT) Subject: [RFC PATCH net-next 2/3] net-memcg: Record pressure level when under pressure Date: Fri, 1 Sep 2023 14:21:27 +0800 Message-Id: <20230901062141.51972-3-wuyun.abel@bytedance.com> X-Mailer: git-send-email 2.37.3 In-Reply-To: <20230901062141.51972-1-wuyun.abel@bytedance.com> References: <20230901062141.51972-1-wuyun.abel@bytedance.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Rspam-User: X-Rspamd-Server: rspam06 X-Rspamd-Queue-Id: AD460120012 X-Stat-Signature: pp1psjbe1c1ngwd8ng1tsqdbn6xtmosj X-HE-Tag: 1693549448-109911 X-HE-Meta: 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 JFb1oake lic47aTbLz1MLPQHvFK4yLMZ6q2mG8kD29J26G4Qfswn9s/pVPtxr3NBqmvd3FH8woi7716k3zAUFZ4Yu//uy3AhuPTWuWNvh2jO70oaW/rvYzcgqNnJDfrFSZ0eV+Lufdl2vbG85hZUVbifBOaaaZXEYHVFImIKNTrg8U9It10u3zyGo0nDDq3X1IcmtyZ64JLSOaEkz8/Vvatz+NW6sIcguihfvCCNwaJDjbxOs8vTw84GRErnWRkXW/gnIT1WUyK1I/tX0F32Sk2cEeVrkklh6OVcQytiyqfY8RMpwd5VegnQHBXwM4AW8MaAnQq+QZQfnRocqfbfqWl/6f6slruy23+AEowPvGmJEl+HHZAwKwgxdjpDxJvofrI8Qt62uNkCnB1B38DVOyzebs7f33AK/1BczGrlCFpJrDAhKYd1I3yPfZw2qS1GbIRGBVaRjBSYSm/lztKk6vu8LbJi9xsXzaicCw3Dh5gMoxyxLPZdD+mc= 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: For now memcg->socket_pressure is used for judging whether there is memory reclaim pressure in this memcg. As different reclaim efficiencies require different strategies, recording the level of pressure would help do fine-grained control inside networking where performance matters a lot. The vmpressure infrastructure classifies pressure into 3 levels: low, medium and critical. It would be too much conservative if constraining socket memory usage at "low" level, so now only the other two are taken into consideration and the least significant bit of socket_pressure is enough to record this information. Signed-off-by: Abel Wu --- include/linux/memcontrol.h | 39 +++++++++++++++++++++++++++++++++----- include/net/sock.h | 2 +- include/net/tcp.h | 2 +- mm/vmpressure.c | 9 ++++++++- 4 files changed, 44 insertions(+), 8 deletions(-) diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index dbf26bc89dd4..a24047bf7722 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -288,6 +288,9 @@ struct mem_cgroup { * Hint of reclaim pressure for socket memroy management. Note * that this indicator should NOT be used in legacy cgroup mode * where socket memory is accounted/charged separately. + * + * The least significant bit is used for indicating the level of + * pressure, 1 for 'critical' and 0 otherwise. */ unsigned long socket_pressure; @@ -1730,15 +1733,40 @@ extern struct static_key_false memcg_sockets_enabled_key; #define mem_cgroup_sockets_enabled static_branch_unlikely(&memcg_sockets_enabled_key) void mem_cgroup_sk_alloc(struct sock *sk); void mem_cgroup_sk_free(struct sock *sk); -static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg) + +static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg, + bool *critical) { + bool under_pressure = false; + + /* + * When cgroup is in legacy mode where tcpmem is separately + * charged, we have no idea about memcg reclaim pressure from + * here and actually no need to, so just ignore the pressure + * level info. + */ if (!cgroup_subsys_on_dfl(memory_cgrp_subsys)) return !!memcg->tcpmem_pressure; + + if (critical) + *critical = false; + do { - if (time_before(jiffies, READ_ONCE(memcg->socket_pressure))) - return true; + unsigned long expire = READ_ONCE(memcg->socket_pressure); + + if (time_before(jiffies, expire)) { + if (!under_pressure) + under_pressure = true; + if (!critical) + break; + if (expire & 1) { + *critical = true; + break; + } + } } while ((memcg = parent_mem_cgroup(memcg))); - return false; + + return under_pressure; } int alloc_shrinker_info(struct mem_cgroup *memcg); @@ -1749,7 +1777,8 @@ void reparent_shrinker_deferred(struct mem_cgroup *memcg); #define mem_cgroup_sockets_enabled 0 static inline void mem_cgroup_sk_alloc(struct sock *sk) { }; static inline void mem_cgroup_sk_free(struct sock *sk) { }; -static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg) +static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg, + bool *critical) { return false; } diff --git a/include/net/sock.h b/include/net/sock.h index 11d503417591..079bbee5c400 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -1434,7 +1434,7 @@ static inline bool sk_under_memory_pressure(const struct sock *sk) return false; if (mem_cgroup_sockets_enabled && sk->sk_memcg && - mem_cgroup_under_socket_pressure(sk->sk_memcg)) + mem_cgroup_under_socket_pressure(sk->sk_memcg, NULL)) return true; return !!READ_ONCE(*sk->sk_prot->memory_pressure); diff --git a/include/net/tcp.h b/include/net/tcp.h index 07b21d9a9620..81e1f9c90a94 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -261,7 +261,7 @@ extern unsigned long tcp_memory_pressure; static inline bool tcp_under_memory_pressure(const struct sock *sk) { if (mem_cgroup_sockets_enabled && sk->sk_memcg && - mem_cgroup_under_socket_pressure(sk->sk_memcg)) + mem_cgroup_under_socket_pressure(sk->sk_memcg, NULL)) return true; return READ_ONCE(tcp_memory_pressure); diff --git a/mm/vmpressure.c b/mm/vmpressure.c index 22c6689d9302..5a3ac3768c0f 100644 --- a/mm/vmpressure.c +++ b/mm/vmpressure.c @@ -308,6 +308,13 @@ void vmpressure(gfp_t gfp, struct mem_cgroup *memcg, bool tree, level = vmpressure_calc_level(scanned, reclaimed); if (level > VMPRESSURE_LOW) { + unsigned long expire = jiffies + HZ; + + if (level == VMPRESSURE_CRITICAL) + expire |= 1UL; + else + expire &= ~1UL; + /* * Let the socket buffer allocator know that * we are having trouble reclaiming LRU pages. @@ -316,7 +323,7 @@ void vmpressure(gfp_t gfp, struct mem_cgroup *memcg, bool tree, * asserted for a second in which subsequent * pressure events can occur. */ - WRITE_ONCE(memcg->socket_pressure, jiffies + HZ); + WRITE_ONCE(memcg->socket_pressure, expire); } } } -- 2.37.3