Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:5 as permitted sender) client-ip=2620:137:e000::3:5;
Date:   Wed, 4 Oct 2023 16:55:11 -0700
From:   Jaegeuk Kim <jaegeuk@kernel.org>
To:     Daeho Jeong <daeho43@gmail.com>
Cc:     linux-kernel@vger.kernel.org,
        linux-f2fs-devel@lists.sourceforge.net, kernel-team@android.com,
        Daeho Jeong <daehojeong@google.com>
Subject: Re: [PATCH] f2fs-tools: use proper address entry count for direct
 nodes
Message-ID: <ZR37Xyr9HcVuLqxx@google.com>
References: <20231003230155.355807-1-daeho43@gmail.com>
 <ZR30mNQEyRjzr0jm@google.com>
 <CACOAw_zSvAPTREb0dSv6srRuxuKxDZp4rMhLyWTk+uJ3WKSWKA@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <CACOAw_zSvAPTREb0dSv6srRuxuKxDZp4rMhLyWTk+uJ3WKSWKA@mail.gmail.com>
Precedence: bulk

On 10/04, Daeho Jeong wrote:
> On Wed, Oct 4, 2023 at 4:26 PM Jaegeuk Kim <jaegeuk@kernel.org> wrote:
> >
> > On 10/03, Daeho Jeong wrote:
> > > From: Daeho Jeong <daehojeong@google.com>
> > >
> > > For direct nodes, we have to use DEF_ADDRS_PER_BLOCK.
> > >
> > > Signed-off-by: Daeho Jeong <daehojeong@google.com>
> > > ---
> > >  fsck/fsck.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/fsck/fsck.c b/fsck/fsck.c
> > > index 78ffdb6..56a7d31 100644
> > > --- a/fsck/fsck.c
> > > +++ b/fsck/fsck.c
> > > @@ -2894,7 +2894,7 @@ static void fsck_failed_reconnect_file_dnode(struct f2fs_sb_info *sbi,
> > >       fsck->chk.valid_blk_cnt--;
> > >       f2fs_clear_main_bitmap(sbi, ni.blk_addr);
> > >
> > > -     for (i = 0; i < ADDRS_PER_BLOCK(&node->i); i++) {
> > > +     for (i = 0; i < DEF_ADDRS_PER_BLOCK; i++) {
> >
> > It seems we need to use the inode block passing by fsck_failed_reconnect_file().
> 
> This function is for direct nodes. Is it correct to use inode block here?

 523 unsigned int addrs_per_block(struct f2fs_inode *i)
 524 {
 525         if (!LINUX_S_ISREG(le16_to_cpu(i->i_mode)) ||
 526                         !(le32_to_cpu(i->i_flags) & F2FS_COMPR_FL))
 527                 return DEF_ADDRS_PER_BLOCK;
 528         return ALIGN_DOWN(DEF_ADDRS_PER_BLOCK, 1 << i->i_log_cluster_size);
 529 }

If the inode is compressed, it seems it has to be aligned to cluster size.

> 
> >
> > >               addr = le32_to_cpu(node->dn.addr[i]);
> > >               if (!addr)
> > >                       continue;
> >
> > 3012                         fsck->chk.valid_blk_cnt--;
> > 3013                         if (addr == NEW_ADDR)
> >
> > And, we also need to skip if addr == COMPRESS_ADDR here?
> >
> > 3014                                 continue;
> > 3015                         f2fs_clear_main_bitmap(sbi, addr);
> > 3016                 }
> >
> > > --
> > > 2.42.0.582.g8ccd20d70d-goog