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 X-Spam-Level: X-Spam-Status: No, score=-8.5 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_HELO_NONE,SPF_PASS, URIBL_BLOCKED,USER_AGENT_SANE_1 autolearn=ham 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 7CA1EC433DF for ; Wed, 1 Jul 2020 07:48:09 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 62FE920722 for ; Wed, 1 Jul 2020 07:48:09 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1728356AbgGAHsI (ORCPT ); Wed, 1 Jul 2020 03:48:08 -0400 Received: from szxga04-in.huawei.com ([45.249.212.190]:7327 "EHLO huawei.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1728294AbgGAHsH (ORCPT ); Wed, 1 Jul 2020 03:48:07 -0400 Received: from DGGEMS412-HUB.china.huawei.com (unknown [172.30.72.60]) by Forcepoint Email with ESMTP id CD71E9B8C42A25890AA8; Wed, 1 Jul 2020 15:48:04 +0800 (CST) Received: from [10.134.22.195] (10.134.22.195) by smtp.huawei.com (10.3.19.212) with Microsoft SMTP Server (TLS) id 14.3.487.0; Wed, 1 Jul 2020 15:48:00 +0800 Subject: Re: [f2fs-dev] [PATCH v3] f2fs: avoid readahead race condition From: Chao Yu To: Jaegeuk Kim , Nathan Chancellor CC: , , References: <20200624012148.180050-1-jaegeuk@kernel.org> <20200629150323.GA3293033@google.com> <20200629202720.GA230664@google.com> <20200630204348.GA2504307@ubuntu-s3-xlarge-x86> <20200630205635.GB1396584@google.com> <285a4e16-2cbc-d1e9-8464-8a06bacbaaa0@huawei.com> Message-ID: Date: Wed, 1 Jul 2020 15:47:59 +0800 User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1 MIME-Version: 1.0 In-Reply-To: <285a4e16-2cbc-d1e9-8464-8a06bacbaaa0@huawei.com> Content-Type: text/plain; charset="utf-8" Content-Language: en-US Content-Transfer-Encoding: 8bit X-Originating-IP: [10.134.22.195] X-CFilter-Loop: Reflected Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 2020/7/1 9:59, Chao Yu wrote: > On 2020/7/1 4:56, Jaegeuk Kim wrote: >> On 06/30, Nathan Chancellor wrote: >>> On Mon, Jun 29, 2020 at 01:27:20PM -0700, Jaegeuk Kim wrote: >>>> If two readahead threads having same offset enter in readpages, every read >>>> IOs are split and issued to the disk which giving lower bandwidth. >>>> >>>> This patch tries to avoid redundant readahead calls. >>>> >>>> Signed-off-by: Jaegeuk Kim >>>> --- >>>> v3: >>>> - use READ|WRITE_ONCE >>>> v2: >>>> - add missing code to bypass read >>>> >>>> fs/f2fs/data.c | 18 ++++++++++++++++++ >>>> fs/f2fs/f2fs.h | 1 + >>>> fs/f2fs/super.c | 2 ++ >>>> 3 files changed, 21 insertions(+) >>>> >>>> diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c >>>> index 995cf78b23c5e..360b4c9080d97 100644 >>>> --- a/fs/f2fs/data.c >>>> +++ b/fs/f2fs/data.c >>>> @@ -2296,6 +2296,7 @@ static int f2fs_mpage_readpages(struct inode *inode, >>>> unsigned nr_pages = rac ? readahead_count(rac) : 1; >>>> unsigned max_nr_pages = nr_pages; >>>> int ret = 0; >>>> + bool drop_ra = false; >>>> >>>> map.m_pblk = 0; >>>> map.m_lblk = 0; >>>> @@ -2306,10 +2307,24 @@ static int f2fs_mpage_readpages(struct inode *inode, >>>> map.m_seg_type = NO_CHECK_TYPE; >>>> map.m_may_create = false; >>>> >>>> + /* >>>> + * Two readahead threads for same address range can cause race condition >>>> + * which fragments sequential read IOs. So let's avoid each other. >>>> + */ >>>> + if (rac && readahead_count(rac)) { >>>> + if (READ_ONCE(F2FS_I(inode)->ra_offset) == readahead_index(rac)) >>>> + drop_ra = true; >>>> + else >>>> + WRITE_ONCE(F2FS_I(inode)->ra_offset, >>>> + readahead_index(rac)); >>>> + } >>>> + >>>> for (; nr_pages; nr_pages--) { >>>> if (rac) { >>>> page = readahead_page(rac); >>>> prefetchw(&page->flags); >>>> + if (drop_ra) >>>> + goto next_page; >>> >>> When CONFIG_F2FS_FS_COMPRESSION is not set (i.e. x86_64 defconfig + >>> CONFIG_F2FS_FS=y): >>> >>> $ make -skj"$(nproc)" O=out distclean defconfig fs/f2fs/data.o >>> ../fs/f2fs/data.c: In function ‘f2fs_mpage_readpages’: >>> ../fs/f2fs/data.c:2327:5: error: label ‘next_page’ used but not defined >>> 2327 | goto next_page; >>> | ^~~~ >>> ... >> >> Thanks. I pushed the fix for -next. >> https://lore.kernel.org/linux-f2fs-devel/1be18397-7fc6-703e-121b-e210e101357f@infradead.org/T/#t It will hang the kernel because we missed to unlock those cached pages, I changed to 'goto set_error_page', the issue was gone. Thanks, > > Reviewed-by: Chao Yu > > Thanks, > >> >> Thanks, >> >>> >>> Cheers, >>> Nathan >>> >>> >>> _______________________________________________ >>> Linux-f2fs-devel mailing list >>> Linux-f2fs-devel@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel >> >> >> _______________________________________________ >> Linux-f2fs-devel mailing list >> Linux-f2fs-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel >> > > > _______________________________________________ > Linux-f2fs-devel mailing list > Linux-f2fs-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel >