Re: [PATCH] ocfs2: kill osb->system_file_mutex lock

[Heming Zhao]

Hello Joseph,

Do you agree that we need to add a rule for the ocfs2 read/write (IO) path?
- When an ocfs2 volume is in the process of mounting, all write
  operations must fail immediately.

- Heming

On 6/24/25 09:55, Tetsuo Handa wrote:
> On 2025/06/24 10:33, Heming Zhao wrote:
> > Hello,
> >
> > Protecting refcnt with a mutex is the right approach, and commit 43b10a20372d
> > did it properly.
> > However, I don't see how your patch fixes the syzbot report [1]. Could you
> > elaborate on the root cause analysis?
> >
> > My review comments are inline below.
> >
> > [1]: https://syzkaller.appspot.com/bug?extid=1fed2de07d8e11a3ec1b
>
> My patch does not fix [1]. My patch fixes a bug which syzbot reported at
> https://lkml.kernel.org/r/000000000000ff2d7a0620381afe@xxxxxxxxxx
> when testing with Diogo's patch at
> https://syzkaller.appspot.com/x/patch.diff?x=178f93d5980000 for [1].
>
> > > Reported-by: Diogo Jahchan Koike <djahchankoike@xxxxxxxxx>
> > 'Reported-by' should be: https://syzkaller.appspot.com/bug?extid=1fed2de07d8e11a3ec1b
>
> Since there is not yet a bug link for my patch, I don't choose syzbot as reporter.
> Diogo will post a formal patch for fixing [1] after returning from vacation.