Re: [PATCH -next 1/4] blk-iocost: track whether iocg is still online

[Yu Kuai <yukuai1@xxxxxxxxxxxxxxx>]

Hi, Christoph

在 2022/12/21 18:33, Christoph Hellwig 写道:
> On Sat, Dec 17, 2022 at 11:05:24AM +0800, Yu Kuai wrote:
> > @@ -459,6 +459,8 @@ struct ioc_gq {
> >   	struct blkg_policy_data		pd;
> >   	struct ioc			*ioc;
> >   
> > +	bool online;
>
> Nit: maybe tab align this field like the fields above it.
>
> > +static void ioc_pd_offline(struct blkg_policy_data *pd)
> > +{
> > +	struct ioc_gq *iocg = pd_to_iocg(pd);
> > +	struct ioc *ioc = iocg->ioc;
> > +	unsigned long flags;
> > +
> > +	if (ioc) {
>
> How could ioc be NULL here?

As I explained in another thread.. pd_offline_fn() can be called without
pd_init_fn(), which is a bug from upper layer...

blkcg_activate_policy
 spin_lock_irq(&q->queue_lock);
 list_for_each_entry_reverse(blkg, &q->blkg_list
  pd_alloc_fn(GFP_NOWAIT | __GFP_NOWARN,...) -> failed

  spin_unlock_irq(&q->queue_lock);
  // release queue_lock here is problematic, this will cause
pd_offline_fn called without pd_init_fn.
  pd_alloc_fn(__GFP_NOWARN,...)

Thanks,
Kuai
> .