MIME-Version: 1.0
References: <20211122030038.1938875-1-cgxu519@mykernel.net> <20211122030038.1938875-8-cgxu519@mykernel.net>
In-Reply-To: <20211122030038.1938875-8-cgxu519@mykernel.net>
From:   Amir Goldstein <amir73il@gmail.com>
Date:   Mon, 22 Nov 2021 09:40:59 +0200
Message-ID: <CAOQ4uxhrg=MAL7sArmP47oyF_QmhG-1b=srs30VNdiT-9s-P0w@mail.gmail.com>
Subject: Re: [RFC PATCH V6 7/7] ovl: implement containerized syncfs for overlayfs
To:     Chengguang Xu <cgxu519@mykernel.net>
Cc:     Miklos Szeredi <miklos@szeredi.hu>, Jan Kara <jack@suse.cz>,
        overlayfs <linux-unionfs@vger.kernel.org>,
        linux-fsdevel <linux-fsdevel@vger.kernel.org>,
        linux-kernel <linux-kernel@vger.kernel.org>,
        Chengguang Xu <charliecgxu@tencent.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, Nov 22, 2021 at 5:01 AM Chengguang Xu <cgxu519@mykernel.net> wrote:
>
> From: Chengguang Xu <charliecgxu@tencent.com>
>
> Now overlayfs can only sync own dirty inodes during syncfs,
> so remove unnecessary sync_filesystem() on upper file system.
>
> Signed-off-by: Chengguang Xu <charliecgxu@tencent.com>
> ---
>  fs/overlayfs/super.c | 14 +++++---------
>  1 file changed, 5 insertions(+), 9 deletions(-)
>
> diff --git a/fs/overlayfs/super.c b/fs/overlayfs/super.c
> index ccffcd96491d..213b795a6a86 100644
> --- a/fs/overlayfs/super.c
> +++ b/fs/overlayfs/super.c
> @@ -292,18 +292,14 @@ static int ovl_sync_fs(struct super_block *sb, int wait)
>         /*
>          * Not called for sync(2) call or an emergency sync (SB_I_SKIP_SYNC).
>          * All the super blocks will be iterated, including upper_sb.
> -        *
> -        * If this is a syncfs(2) call, then we do need to call
> -        * sync_filesystem() on upper_sb, but enough if we do it when being
> -        * called with wait == 1.
>          */
> -       if (!wait)
> -               return 0;
> -
>         upper_sb = ovl_upper_mnt(ofs)->mnt_sb;
> -
>         down_read(&upper_sb->s_umount);
> -       ret = sync_filesystem(upper_sb);
> +       if (wait)
> +               wait_sb_inodes(upper_sb);
> +       if (upper_sb->s_op->sync_fs)
> +               upper_sb->s_op->sync_fs(upper_sb, wait);
> +       ret = ovl_sync_upper_blockdev(upper_sb, wait);

I think it will be cleaner to use a helper ovl_sync_upper_filesystem()
with everything from  upper_sb = ... and a comment to explain that
this is a variant of __sync_filesystem() where all the dirty inodes write
have already been started.

Thanks,
Amir.

P.S. I like this "stoopid proof" v6 because I can understand it ;-)