Mat Martineau <mathew.j.martineau(a)linux.intel.com> wrote:
> > -		pr_debug("parent=%p", parent);
> > +		struct socket_wq *wq;
> > 
> > 		smp_mb__before_atomic();
> > 		set_bit(MPTCP_DATA_READY, &mptcp_sk(parent)->flags);
> > 		smp_mb__after_atomic();
> > 
> > -		parent->sk_data_ready(parent);
> > -	}
> > +		rcu_read_lock();
> > +		wq = rcu_dereference(parent->sk_wq);
> > +		if (skwq_has_sleeper(wq) &&
> > +		    mptcp_should_wake_parent(subflow)) {
> > +			wake_up_interruptible_all(&wq->wait);
> > +		}
> > +		rcu_read_unlock();
> 
> Compared to sock_def_readable, this wakeup code doesn't call sk_wake_async.
> Is SOCK_FASYNC irrelevant here (sk_wake_async doesn't do anything without
> that sock flag)?

Its not.  Perhaps best to make this a simpler

if (mptcp_should_wake_parent(subflow))
	parent->sk_data_ready(parent);

I will make this change in the next iteration.