Received: by 2002:a05:6902:102b:0:0:0:0 with SMTP id x11csp453892ybt; Wed, 17 Jun 2020 05:33:45 -0700 (PDT) X-Google-Smtp-Source: ABdhPJz4cAPqCQK+gemwuoauvUCNsEmD2uj6AgN8U9J6KuYUkHKOLnma1WBsDKucFur9EF4JIfz8 X-Received: by 2002:a17:906:4554:: with SMTP id s20mr7092288ejq.241.1592397225357; Wed, 17 Jun 2020 05:33:45 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1592397225; cv=none; d=google.com; s=arc-20160816; b=aofIv2lTubV2WicHHiWl//Vei83nsIIPIfq0FybB3oFqEM4lKt85DCoOZrM1KhEH7u 1CDHmaPjK7dw6/eWskgnYkALbtR/xezrPobMBBZfR8KcKGZsLeGerdwZSatwGPE5oECS LN6LOFvANWmEH6EOfNnS9T1gQiP+cHF3K+QCYvEmdlGtGrlva8eSdHC64WpRX32BPZpq Y30KWUhXcDI1lLgSdeJRXenUNQDLstpM+d+Dym4MJDQRNKykxB0cZFcP+XJFD6dOb080 dRj7dNj4p82iJpdLgticY7IBLLxgEO4WbSdDDgG2rCyo34L9g7P2tIVnTM4qpCoakaBI QXcA== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:sender:content-transfer-encoding :content-language:in-reply-to:mime-version:user-agent:date :message-id:from:references:cc:to:subject; bh=nFsdFWcbLYBUlCYK3kE4wKXFEiCnwdJ9C5Nc4wdtPBk=; b=KX84truBKLHc8qvvOYn1EBJR+C2B6W646DWd5pC7uOxEGMJjyOYsyRDwXMsSTUzSX2 h9XWr4HgYA8n7GByWlcdAa/EWasrmEDnyB3vJoF5QpIkfkX4Ua0M+W+CSHO1kTeM80fj xUMv1cnMmqHg2HscZgOuj1s1hVu8c5esgFxqDYVrf9jqfGCVupUVG6NP6dajlVb38qjx J9ioDItY1i5vrukg0/2s6duj9z2L7sPLyt98I1mlpqBX6viCRBzJ2gJA/9z6qCRjYCf+ CqBNY8ctS1/GElROfTlIvgIkEWKcN4mJyqJW6FkBTlpj9FLosrzEHx8N7gfyzILVgUox jy6g== ARC-Authentication-Results: i=1; mx.google.com; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Return-Path: Received: from vger.kernel.org (vger.kernel.org. [23.128.96.18]) by mx.google.com with ESMTP id k7si13407176ejr.30.2020.06.17.05.33.22; Wed, 17 Jun 2020 05:33:45 -0700 (PDT) Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18; Authentication-Results: mx.google.com; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726702AbgFQM20 (ORCPT + 99 others); Wed, 17 Jun 2020 08:28:26 -0400 Received: from szxga07-in.huawei.com ([45.249.212.35]:39652 "EHLO huawei.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1726308AbgFQM2Z (ORCPT ); Wed, 17 Jun 2020 08:28:25 -0400 Received: from DGGEMS414-HUB.china.huawei.com (unknown [172.30.72.60]) by Forcepoint Email with ESMTP id 7EC3826641B5BDBD1DA9; Wed, 17 Jun 2020 20:28:21 +0800 (CST) Received: from [10.134.22.195] (10.134.22.195) by smtp.huawei.com (10.3.19.214) with Microsoft SMTP Server (TLS) id 14.3.487.0; Wed, 17 Jun 2020 20:28:16 +0800 Subject: Re: [PATCH] f2fs: fix a race condition between f2fs_write_end_io and f2fs_del_fsync_node_entry To: , CC: , References: <1592384659-20203-1-git-send-email-zhaowuyun@wingtech.com> From: Chao Yu Message-ID: <86069ba3-cba3-7bc9-c90b-e931abd0dde5@huawei.com> Date: Wed, 17 Jun 2020 20:28:16 +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: <1592384659-20203-1-git-send-email-zhaowuyun@wingtech.com> Content-Type: text/plain; charset="windows-1252" Content-Language: en-US Content-Transfer-Encoding: 7bit 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 2020/6/17 17:04, zhaowuyun@wingtech.com wrote: > From: Wuyun Zhao > > Under some condition, the __write_node_page will submit a page which is not > f2fs_in_warm_node_list and will not call f2fs_add_fsync_node_entry. > f2fs_gc continue to run to invoke f2fs_iget -> do_read_inode to read the same node page > and set code node, which make f2fs_in_warm_node_list become true, > that will cause f2fs_bug_on in f2fs_del_fsync_node_entry when f2fs_write_end_io called. Could you please add below race condition description into commit message? - f2fs_write_end_io - f2fs_iget - do_read_inode - set_cold_node recover cold node flag - f2fs_in_warm_node_list - is_cold_node if node is cold, assume we have added node to fsync_node_list during writepages() - f2fs_del_fsync_node_entry - f2fs_bug_on() due to node page is not in fsync_node_list BTW, I'm curious about why we can lose cold flag for non-dir inode? any clue to reproduce this bug (I mean losing cold flag)? > > [ 34.966133] Call trace: > [ 34.969902] f2fs_del_fsync_node_entry+0x100/0x108 > [ 34.976071] f2fs_write_end_io+0x1e0/0x288 > [ 34.981539] bio_endio+0x248/0x270 > [ 34.986289] blk_update_request+0x2b0/0x4d8 > [ 34.991841] scsi_end_request+0x40/0x440 > [ 34.997126] scsi_io_completion+0xa4/0x748 > [ 35.002593] scsi_finish_command+0xdc/0x110 > [ 35.008143] scsi_softirq_done+0x118/0x150 > [ 35.013610] blk_done_softirq+0x8c/0xe8 > [ 35.018811] __do_softirq+0x2e8/0x578 > [ 35.023828] irq_exit+0xfc/0x120 > [ 35.028398] handle_IPI+0x1d8/0x330 > [ 35.033233] gic_handle_irq+0x110/0x1d4 > [ 35.038433] el1_irq+0xb4/0x130 > [ 35.042917] kmem_cache_alloc+0x3f0/0x418 > [ 35.048288] radix_tree_node_alloc+0x50/0xf8 > [ 35.053933] __radix_tree_create+0xf8/0x188 > [ 35.059484] __radix_tree_insert+0x3c/0x128 > [ 35.065035] add_gc_inode+0x90/0x118 > [ 35.069967] f2fs_gc+0x1b80/0x2d70 > [ 35.074718] f2fs_disable_checkpoint+0x94/0x1d0 > [ 35.080621] f2fs_fill_super+0x10c4/0x1b88 > [ 35.086088] mount_bdev+0x194/0x1e0 > [ 35.090923] f2fs_mount+0x40/0x50 > [ 35.095589] mount_fs+0xb4/0x190 > [ 35.100159] vfs_kern_mount+0x80/0x1d8 > [ 35.105260] do_mount+0x478/0xf18 > [ 35.109926] ksys_mount+0x90/0xd0 > [ 35.114592] __arm64_sys_mount+0x24/0x38 > > Signed-off-by: Wuyun Zhao Reviewed-by: Chao Yu Thanks, > --- > fs/f2fs/inode.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c > index be6ac33..0df5c8c 100644 > --- a/fs/f2fs/inode.c > +++ b/fs/f2fs/inode.c > @@ -402,6 +402,7 @@ static int do_read_inode(struct inode *inode) > > /* try to recover cold bit for non-dir inode */ > if (!S_ISDIR(inode->i_mode) && !is_cold_node(node_page)) { > + f2fs_wait_on_page_writeback(node_page, NODE, true, true); > set_cold_node(node_page, false); > set_page_dirty(node_page); > } >