From patchwork Thu Dec 10 08:08:43 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: SeongJae Park X-Patchwork-Id: 342691 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-13.8 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH, MAILING_LIST_MULTI, SPF_HELO_NONE, SPF_PASS, USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id B60A8C433FE for ; Thu, 10 Dec 2020 08:10:13 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 85DC823D51 for ; Thu, 10 Dec 2020 08:10:13 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1733296AbgLJIJ5 (ORCPT ); Thu, 10 Dec 2020 03:09:57 -0500 Received: from smtp-fw-6001.amazon.com ([52.95.48.154]:58786 "EHLO smtp-fw-6001.amazon.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1725896AbgLJIJ5 (ORCPT ); Thu, 10 Dec 2020 03:09:57 -0500 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=amazon.com; i=@amazon.com; q=dns/txt; s=amazon201209; t=1607587796; x=1639123796; h=from:to:cc:subject:date:message-id:mime-version; bh=aEJzy3+FKb9u4xePy0sLXdWn5xOn3WQ9Rfk09qhrQYk=; b=KEdZGycvm1b326PFi92VZyT9la9OZgCStrhOvoboi7XEVNSMPWHE+jiK JV94u/VUvz6DYPqXNOTUky3MpGsjZD/zbBF2h6Q39xL/DyyQoAVIqqG60 ztv3R93F0HmqtRUHzZ3ZxJAUT0sO0fg8tyPc5bKaUdcMrOK2Su/dI48wW 0=; X-IronPort-AV: E=Sophos;i="5.78,407,1599523200"; d="scan'208";a="71701030" Received: from iad12-co-svc-p1-lb1-vlan2.amazon.com (HELO email-inbound-relay-1a-af6a10df.us-east-1.amazon.com) ([10.43.8.2]) by smtp-border-fw-out-6001.iad6.amazon.com with ESMTP; 10 Dec 2020 08:09:09 +0000 Received: from EX13D31EUA001.ant.amazon.com (iad12-ws-svc-p26-lb9-vlan3.iad.amazon.com [10.40.163.38]) by email-inbound-relay-1a-af6a10df.us-east-1.amazon.com (Postfix) with ESMTPS id 6BDBFA18C4; Thu, 10 Dec 2020 08:09:07 +0000 (UTC) Received: from u3f2cd687b01c55.ant.amazon.com (10.43.161.102) by EX13D31EUA001.ant.amazon.com (10.43.165.15) with Microsoft SMTP Server (TLS) id 15.0.1497.2; Thu, 10 Dec 2020 08:09:02 +0000 From: SeongJae Park To: CC: SeongJae Park , , , , , , , , Subject: [PATCH v2 0/1] net: Reduce rcu_barrier() contentions from 'unshare(CLONE_NEWNET)' Date: Thu, 10 Dec 2020 09:08:43 +0100 Message-ID: <20201210080844.23741-1-sjpark@amazon.com> X-Mailer: git-send-email 2.17.1 MIME-Version: 1.0 X-Originating-IP: [10.43.161.102] X-ClientProxiedBy: EX13D16UWB004.ant.amazon.com (10.43.161.170) To EX13D31EUA001.ant.amazon.com (10.43.165.15) Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org From: SeongJae Park On a few of our systems, I found frequent 'unshare(CLONE_NEWNET)' calls make the number of active slab objects including 'sock_inode_cache' type rapidly and continuously increase. As a result, memory pressure occurs. In more detail, I made an artificial reproducer that resembles the workload that we found the problem and reproduce the problem faster. It merely repeats 'unshare(CLONE_NEWNET)' 50,000 times in a loop. It takes about 2 minutes. On 40 CPU cores, 70GB DRAM machine, it reduced about 15GB of available memory in total. Note that the issue don't reproduce on every machine. On my 6 CPU cores machine, the problem didn't reproduce. 'cleanup_net()' and 'fqdir_work_fn()' are functions that deallocate the relevant memory objects. They are asynchronously invoked by the work queues and internally use 'rcu_barrier()' to ensure safe destructions. 'cleanup_net()' works in a batched maneer in a single thread worker, while 'fqdir_work_fn()' works for each 'fqdir_exit()' call in the 'system_wq'. Therefore, 'fqdir_work_fn()' called frequently under the workload and made the contention for 'rcu_barrier()' high. In more detail, the global mutex, 'rcu_state.barrier_mutex' became the bottleneck. I tried making 'fqdir_work_fn()' batched and confirmed it works. The following patch is for the change. I think this is the right solution for point fix of this issue, but someone might blame different parts. 1. User: Frequent 'unshare()' calls