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;
Date:   Thu, 7 May 2020 15:23:58 +0800
From:   Gao Xiang <hsiangkao@gmx.com>
To:     Chao Yu <yuchao0@huawei.com>
Cc:     Eric Biggers <ebiggers@kernel.org>,
        Jaegeuk Kim <jaegeuk@kernel.org>, kernel-team@android.com,
        linux-kernel@vger.kernel.org,
        linux-f2fs-devel@lists.sourceforge.net
Subject: Re: [f2fs-dev] [PATCH] f2fs: get parent inode when recovering pino
Message-ID: <20200507072349.GA5471@hsiangkao-HP-ZHAN-66-Pro-G1>
References: <20200505165847.GA98848@gmail.com>
 <20200505181323.GA55221@google.com>
 <20200505181941.GC98848@gmail.com>
 <20200506001403.GA2101@hsiangkao-HP-ZHAN-66-Pro-G1>
 <20200506012428.GG128280@sol.localdomain>
 <20200506015813.GA9256@hsiangkao-HP-ZHAN-66-Pro-G1>
 <20200506064709.GA25482@hsiangkao-HP-ZHAN-66-Pro-G1>
 <20200506191613.GB842@sol.localdomain>
 <20200506223623.GA27760@hsiangkao-HP-ZHAN-66-Pro-G1>
 <3a837113-8081-6f01-4d8d-1d4b6600ec8c@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <3a837113-8081-6f01-4d8d-1d4b6600ec8c@huawei.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Chao,

On Thu, May 07, 2020 at 02:38:39PM +0800, Chao Yu wrote:
> On 2020/5/7 6:36, Gao Xiang wrote:
> > On Wed, May 06, 2020 at 12:16:13PM -0700, Eric Biggers wrote:
> >> On Wed, May 06, 2020 at 02:47:19PM +0800, Gao Xiang wrote:
> >>> On Wed, May 06, 2020 at 09:58:22AM +0800, Gao Xiang wrote:
> >>>> On Tue, May 05, 2020 at 06:24:28PM -0700, Eric Biggers wrote:
> >>>>> On Wed, May 06, 2020 at 08:14:07AM +0800, Gao Xiang wrote:
> >>>>>>>
> >>>>>>> Actually, I think this is wrong because the fsync can be done vi=
a a file
> >>>>>>> descriptor that was opened to a now-deleted link to the file.
> >>>>>>
> >>>>>> I'm still confused about this...
> >>>>>>
> >>>>>> I don't know what's wrong with this version from my limited knowl=
edge?
> >>>>>>  inode itself is locked when fsyncing, so
> >>>>>>
> >>>>>>    if the fsync inode->i_nlink =3D=3D 1, this inode has only one =
hard link
> >>>>>>    (not deleted yet) and should belong to a single directory; and
> >>>>>>
> >>>>>>    the only one parent directory would not go away (not deleted a=
s well)
> >>>>>>    since there are some dirents in it (not empty).
> >>>>>>
> >>>>>> Could kindly explain more so I would learn more about this scenar=
io?
> >>>>>> Thanks a lot!
> >>>>>
> >>>>> i_nlink =3D=3D 1 just means that there is one non-deleted link.  T=
here can be links
> >>>>> that have since been deleted, and file descriptors can still be op=
en to them.
> >>>>
> >>>> Thanks for your inspiration. You are right, thanks.
> >>>>
> >>>> Correct my words... I didn't check f2fs code just now, it seems f2f=
s doesn't
> >>>> take inode_lock as some other fs like __generic_file_fsync or ubifs=
_fsync.
> >>>>
> >>>> And i_sem locks nlink / try_to_fix_pino similarly in some extent. I=
t seems
> >>>> no race by using d_find_alias here. Thanks again.
> >>>>
> >>>
> >>> (think more little bit just now...)
> >>>
> >>>  Thread 1:                                           Thread 2 (fsync=
):
> >>>   vfs_unlink                                          try_to_fix_pin=
o
> >>>     f2fs_unlink
> >>>        f2fs_delete_entry
> >>>          f2fs_drop_nlink  (i_sem, inode->i_nlink =3D 1)
> >>>
> >>>   (...   but this dentry still hashed)                  i_sem, check=
 inode->i_nlink =3D 1
> >>>                                                         i_sem d_find=
_alias
> >>>
> >>>   d_delete
> >>>
> >>> I'm not sure if fsync could still use some wrong alias by chance..
> >>> completely untested, maybe just noise...
>
> Another race condition could be:
>
> Thread 1 (fsync)		Thread 2 (rename)
> - f2fs_sync_fs
> - try_to_fix_pino
> 				- f2fs_rename
> 				 - down_write
> 				 - file_lost_pino
> 				 - up_write
>  - down_write
>  - file_got_pino
>  - up_write

Yes, IMHO, I think it could be not proper to
take dir lock in fsync path anyway...

I would suggest as before (if it needs to be fixed).
And it seems no significant performance difference.

Thanks,
Gao Xiang

>
> Thanks,
>
> >>>
> >>
> >> Right, good observation.  My patch makes it better, but it's still br=
oken.
> >>
> >> I don't know how to fix it.  If we see i_nlink =3D=3D 1 and multiple =
hashed
> >> dentries, there doesn't appear to be a way to distingush which one co=
rresponds
> >> to the remaining link on-disk (if any; it may not even be in the dcac=
he), and
> >> which correspond to links that vfs_unlink() has deleted from disk but=
 hasn't yet
> >> done d_delete() on.
> >>
> >> One idea would be choose one, then take inode_lock_shared(dir) and do
> >> __f2fs_find_entry() to check if the dentry is really still on-disk.  =
That's
> >> heavyweight and error-prone though, and the locking could cause probl=
ems.
> >>
> >> I'm wondering though, does f2fs really need try_to_fix_pino() at all,=
 and did it
> >> ever really work?  It never actually updates the f2fs_inode::i_name t=
o match the
> >> new directory.  So independently of this bug with deleted links, I do=
n't see how
> >> it can possibly work as intended.
> >
> > Part of my humble opinion would be "update pino in rename/unlink/link.=
.. such ops
> > instead of in fsync" (maybe it makes better sense of locking)... But a=
ctually I'm
> > not a f2fs folk now, just curious about what the original patch resolv=
ed with
> > these new extra igrab/iput (as I said before, I could not find some cl=
ue previously).
> >
> > Thanks,
> > Gao Xiang
> >
> >>
> >> - Eric
> >
> >
> > _______________________________________________
> > Linux-f2fs-devel mailing list
> > Linux-f2fs-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
> > .
> >