From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753198AbeCUQuq (ORCPT ); Wed, 21 Mar 2018 12:50:46 -0400 Received: from out30-130.freemail.mail.aliyun.com ([115.124.30.130]:34682 "EHLO out30-130.freemail.mail.aliyun.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751880AbeCUQun (ORCPT ); Wed, 21 Mar 2018 12:50:43 -0400 X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R141e4;CH=green;FP=0|-1|-1|-1|0|-1|-1|-1;HT=e01e01429;MF=yang.shi@linux.alibaba.com;NM=1;PH=DS;RN=4;SR=0;TI=SMTPD_---0SzrTt7Q_1521651036; Subject: Re: [RFC PATCH 1/8] mm: mmap: unmap large mapping by section To: Michal Hocko Cc: akpm@linux-foundation.org, linux-mm@kvack.org, linux-kernel@vger.kernel.org References: <1521581486-99134-1-git-send-email-yang.shi@linux.alibaba.com> <1521581486-99134-2-git-send-email-yang.shi@linux.alibaba.com> <20180321131449.GN23100@dhcp22.suse.cz> From: Yang Shi Message-ID: <8e0ded7b-4be4-fa25-f40c-d3116a6db4db@linux.alibaba.com> Date: Wed, 21 Mar 2018 09:50:31 -0700 User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.12; rv:52.0) Gecko/20100101 Thunderbird/52.2.1 MIME-Version: 1.0 In-Reply-To: <20180321131449.GN23100@dhcp22.suse.cz> Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit Content-Language: en-US Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 3/21/18 6:14 AM, Michal Hocko wrote: > On Wed 21-03-18 05:31:19, Yang Shi wrote: >> When running some mmap/munmap scalability tests with large memory (i.e. >>> 300GB), the below hung task issue may happen occasionally. >> INFO: task ps:14018 blocked for more than 120 seconds. >> Tainted: G E 4.9.79-009.ali3000.alios7.x86_64 #1 >> "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this >> message. >> ps D 0 14018 1 0x00000004 >> ffff885582f84000 ffff885e8682f000 ffff880972943000 ffff885ebf499bc0 >> ffff8828ee120000 ffffc900349bfca8 ffffffff817154d0 0000000000000040 >> 00ffffff812f872a ffff885ebf499bc0 024000d000948300 ffff880972943000 >> Call Trace: >> [] ? __schedule+0x250/0x730 >> [] schedule+0x36/0x80 >> [] rwsem_down_read_failed+0xf0/0x150 >> [] call_rwsem_down_read_failed+0x18/0x30 >> [] down_read+0x20/0x40 >> [] proc_pid_cmdline_read+0xd9/0x4e0 > Slightly off-topic: > Btw. this sucks as well. Do we really need to take mmap_sem here? Do any > of > arg_start = mm->arg_start; > arg_end = mm->arg_end; > env_start = mm->env_start; > env_end = mm->env_end; > > change after exec or while the pid is already visible in proc? If yes > maybe we can use a dedicated lock. Actually, Alexey Dobriyan had the same comment when he reviewed my very first patch (which changes down_read to down_read_killable at that place). Those 4 values might be changed by prctl_set_mm() and prctl_set_mm_map() concurrently. They used to use down_read() to protect the change, but it looks not good enough to protect concurrent writing. So, Mateusz Guzik's commit ddf1d398e517e660207e2c807f76a90df543a217 ("prctl: take mmap sem for writing to protect against others") change it to down_write(). It seems mmap_sem can be replaced to a dedicated lock. How about defining a rwlock in mm_struct to protect those data? I will come up with a RFC patch for this. However, this dedicated lock just can work around this specific case. I believe solving mmap_sem scalability issue aimed by the patch series is still our consensus. Thanks, Yang