Re: [External] Re: [PATCH v6] memcg: Don't wait writeback completion when release memcg.

[Julian Sun <sunjunchao@bytedance.com>]

Hi,

On Thu, Sep 18, 2025 at 5:50 AM Tejun Heo <tj@kernel.org> wrote:
>
> Hello,
>
> On Thu, Sep 18, 2025 at 05:29:59AM +0800, Julian Sun wrote:
> ...
> > The wb_wait_for_completion() here is probably only used to prevent
> > use-after-free. Therefore, we manage 'done' separately and automatically
> > free it.
> >
> > This allows us to remove wb_wait_for_completion() while preventing
> > the use-after-free issue.
> >
> > Fixes: 97b27821b485 ("writeback, memcg: Implement foreign dirty flushing")
> > Tested-by: Venkat Rao Bagalkote <venkat88@linux.ibm.com>
> > Signed-off-by: Julian Sun <sunjunchao@bytedance.com>
>
>   Acked-by: Tejun Heo <tj@kernel.org>
>
> Minor comments below:
>
> > +/*
> > + * This structure exists to avoid waiting for writeback to finish on
> > + * memcg release, which could lead to a hang task.
>                                            ^
>                                            hung
>
> > + * @done.cnt is always > 0 before a memcg is released, so @wq_entry.func
> > + * may only be invoked by finish_writeback_work() after memcg is freed.
> > + * See mem_cgroup_css_free() for details.
> > + */
>
> I'm not sure this gives enough of a picture of what's going on. It'd be
> better to expand a bit - briefly describe what the whole mechanism is for
> and how hung tasks can happen.

Ah.. I think explaining how hungtask occurs is a bit too detailed and
verbose. If readers want to learn about more detailed content and
causes, I think the commit message would be a better choice...
>
> Thanks.
>
> --
> tejun


Thanks,
-- 
Julian Sun <sunjunchao@bytedance.com>