Mime-Version: 1.0
Subject: Re: [PATCH] f2fs: handle newly created page when revoking inmem
 pages
Reply-To: daeho.jeong@samsung.com
From: Daeho Jeong <daeho.jeong@samsung.com>
To: "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "linux-f2fs-devel@lists.sourceforge.net" 
        <linux-f2fs-devel@lists.sourceforge.net>
In-Reply-To: <aec4e374-a245-590e-8350-d1e1ada3b705@huawei.com>
Message-ID: <20180110072203epcms1p5590a9b8441e94174b461172222424455@epcms1p5>
Date: Wed, 10 Jan 2018 07:22:03 +0000
Content-Transfer-Encoding: 8BIT
Content-Type: text/plain; charset="utf-8"
CMS-TYPE: 101P
DLP-Filter: Pass
References: <aec4e374-a245-590e-8350-d1e1ada3b705@huawei.com>
        <1515552234-27964-1-git-send-email-daeho.jeong@samsung.com>
        <CGME20180110024336epcas2p1003559800e7f1e59a79496fc5b79cf04@epcms1p5>
Sender: linux-kernel-owner@vger.kernel.org

Hi Chao,

> Original intention here is to recover status to the timing before
> committing atomic write. As at that timing blkaddr in dnode should be
> cur->old_addr(NEW_ADDR), so we need to change to call:
 
> f2fs_update_data_blkaddr(&dn, NEW_ADDR);

Ok, I'll change NULL_ADDR to NEW_ADDR.

Thanks,
 
> Otherwise, metadata will become inconsistent, because blkaddr value is
> NULL_ADDR means that current block is not preallocated, but
> total_valid_block_count has already been updated. Right?
 
> Thanks,