From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jaegeuk Kim Subject: Re: [f2fs-dev] [PATCH 01/10] f2fs: reassign new segment for mode=lfs Date: Wed, 4 Jan 2017 14:48:27 -0800 Message-ID: <20170104224827.GB1011@jaegeuk.local> References: <20161230185117.3832-1-jaegeuk@kernel.org> <702b01ba-5fcb-01a2-0bad-65ea9005b834@huawei.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <702b01ba-5fcb-01a2-0bad-65ea9005b834@huawei.com> Sender: linux-fsdevel-owner@vger.kernel.org To: Chao Yu Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net List-Id: linux-f2fs-devel.lists.sourceforge.net On 01/04, Chao Yu wrote: > Hi Jaegeuk, > > On 2016/12/31 2:51, Jaegeuk Kim wrote: > > Otherwise we can remain wrong curseg->next_blkoff, resulting in fsck failure. > > Could you explain more about this case? I remember that I hit an fsck failure when I was testing f2fs with an smr drive. I didn't dig into the error, but the fact is that our roll-forward recovery doesn't update current segment information at every time, but allocate a new section at the end of the work like below. I just enabled it for the LFS mode in order to avoid that failure. Thanks, > > Thanks, > > > > > Signed-off-by: Jaegeuk Kim > > --- > > fs/f2fs/segment.c | 3 --- > > 1 file changed, 3 deletions(-) > > > > diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c > > index be9e4d244d75..4e5ffe1d97e4 100644 > > --- a/fs/f2fs/segment.c > > +++ b/fs/f2fs/segment.c > > @@ -1428,9 +1428,6 @@ void allocate_new_segments(struct f2fs_sb_info *sbi) > > unsigned int old_segno; > > int i; > > > > - if (test_opt(sbi, LFS)) > > - return; > > - > > for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++) { > > curseg = CURSEG_I(sbi, i); > > old_segno = curseg->segno; > >