Re: [PATCH v3] kernel/fork: beware of __put_task_struct calling context

[Oleg Nesterov]

On 02/02, Wander Lairson Costa wrote:
>
> > We already have delayed_put_task_struct() which differs very much.
> > Perhaps something like ___put_task_struct() will look less confusing.
> >
>
> ___put_task_struct()? I already added a function with this name below.

Ah, I meant ___put_task_struct_rcu() or something like this. Bug again
this is just cosmetic nit, please ignore

> > > +void __put_task_struct(struct task_struct *tsk)
> > > +{
> > > +     if (IS_ENABLED(CONFIG_PREEMPT_RT) && (!preemptible() || !in_task()))
> > > +             /*
> > > +              * under PREEMPT_RT, we can't call put_task_struct
> > > +              * in atomic context because it will indirectly
> > > +              * acquire sleeping locks.
> > > +              */
> > > +             call_rcu(&tsk->rcu, __delayed_put_task_struct);
> >
> > Perhaps this deserves additional note to explain why is it safe to use tsk->rcu
> > union. May be this is obvious, but I was confused when I looked at the previous
> > version ;)
> >
>
> Makes sense, I will add it in the next version.

Thanks ;)

Oleg.