Re: [PATCH 2/3] jump_labels: do not use unserialized static_key_enabled

[Paolo Bonzini]

----- Original Message -----
> From: "Eric Dumazet" <eric.dumazet@xxxxxxxxx>
> To: "Paolo Bonzini" <pbonzini@xxxxxxxxxx>
> Cc: linux-kernel@xxxxxxxxxxxxxxx, "Peter Zijlstra" <peterz@xxxxxxxxxxxxx>
> Sent: Tuesday, August 1, 2017 10:52:14 PM
> Subject: Re: [PATCH 2/3] jump_labels: do not use unserialized static_key_enabled
> 
> On Tue, 2017-08-01 at 17:24 +0200, Paolo Bonzini wrote:
> > Any use of key->enabled (that is static_key_enabled and static_key_count)
> > outside jump_label_lock should handle its own serialization.  The only
> > two that are not doing so are the UDP encapsulation static keys.  Change
> > them to use static_key_enable, which now correctly tests key->enabled under
> > the jump label lock.
> >
> > diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
> > index 1d6219bf2d6b..74b7810df9fc 100644
> > --- a/net/ipv4/udp.c
> > +++ b/net/ipv4/udp.c
> > @@ -1644,8 +1644,7 @@ static int __udp_queue_rcv_skb(struct sock *sk,
> > struct sk_buff *skb)
> >  static struct static_key udp_encap_needed __read_mostly;
> >  void udp_encap_enable(void)
> >  {
> > -	if (!static_key_enabled(&udp_encap_needed))
> > -		static_key_slow_inc(&udp_encap_needed);
> > +	static_key_enable(&udp_encap_needed);
> >  }
> 
> Looks good to me, but static_key_enable() is not serialized either ?
> 
> I suspect you should have CCed me on patch 1/3 :)

Yes, indeed...  That's the "now correctly test" in the commit message.
What a difference one word makes. :)

You can find 1/3 at https://patchwork.kernel.org/patch/9874845/.

Paolo