[RFC] lockdep warning in gen_kill_estimator

On Wed, 08 Sep 2010 20:40:34 +0200
Eric Dumazet <eric.dumazet@gmail.com> wrote:

> Le mercredi 08 septembre 2010 à 11:27 -0700, Stephen Hemminger a écrit :
> > Saw big lockdep SOFTIRQ-unsafe lock warning in gen_kill_estimator when doing.
> > 
> >   tc qdisc del dev dummy0 root
> > 
> 
> > 
> > --- a/net/core/gen_estimator.c	2010-09-08 11:16:10.184368926 -0700
> > +++ b/net/core/gen_estimator.c	2010-09-08 11:16:24.694194050 -0700
> > @@ -270,18 +270,18 @@ void gen_kill_estimator(struct gnet_stat
> >  {
> >  	struct gen_estimator *e;
> >  
> > -	spin_lock(&est_tree_lock);
> > +	spin_lock_bh(&est_tree_lock);
> >  	while ((e = gen_find_node(bstats, rate_est))) {
> >  		rb_erase(&e->node, &est_root);
> >  
> > -		write_lock_bh(&est_lock);
> > +		write_lock(&est_lock);
> >  		e->bstats = NULL;
> > -		write_unlock_bh(&est_lock);
> > +		write_unlock(&est_lock);
> >  
> >  		list_del_rcu(&e->list);
> >  		call_rcu(&e->e_rcu, __gen_kill_estimator);
> >  	}
> > -	spin_unlock(&est_tree_lock);
> > +	spin_unlock_bh(&est_tree_lock);
> >  }
> >  EXPORT_SYMBOL(gen_kill_estimator);
> >  
> 
> Hmm... I think commit 0b5d404e349c0236
> already takes care of this ?
> 
> Already in Linus tree 
> 
> 
Ok.

But that commit adds unnecessary bh_disable
-------
Subject: [PATCH] pkt_sched: remov unnecessary bh_disable

Now that est_tree_lock is acquired with BH protection, the other
call is unnecessary.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>

Le mercredi 08 septembre 2010 à 12:16 -0700, Stephen Hemminger a écrit :

> But that commit adds unnecessary bh_disable
> -------

Yes I mentioned it in the original patch discussion


> Subject: [PATCH] pkt_sched: remov unnecessary bh_disable
> 
> Now that est_tree_lock is acquired with BH protection, the other
> call is unnecessary.
> 
> Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
> 
> --- a/net/core/gen_estimator.c	2010-09-08 12:13:55.182851300 -0700
> +++ b/net/core/gen_estimator.c	2010-09-08 12:14:07.401063663 -0700
> @@ -274,9 +274,9 @@ void gen_kill_estimator(struct gnet_stat
>  	while ((e = gen_find_node(bstats, rate_est))) {
>  		rb_erase(&e->node, &est_root);
>  
> -		write_lock_bh(&est_lock);
> +		write_lock(&est_lock);
>  		e->bstats = NULL;
> -		write_unlock_bh(&est_lock);
> +		write_unlock(&est_lock);
>  
>  		list_del_rcu(&e->list);
>  		call_rcu(&e->e_rcu, __gen_kill_estimator);
> 
> 
> 


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Wed, 08 Sep 2010 21:23:03 +0200

> Le mercredi 08 septembre 2010 à 12:16 -0700, Stephen Hemminger a écrit :
> 
>> But that commit adds unnecessary bh_disable
>> -------
> 
> Yes I mentioned it in the original patch discussion

I'll apply this to net-next-2.6 after my next merge of net-2.6
into net-next-2.6
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html