netdev - Re: [Patch net-next] act_mirred: get rid of mirred_list

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1512083524.19682.29.camel@gmail.com>
Date:   Thu, 30 Nov 2017 15:12:04 -0800
From:   Eric Dumazet <eric.dumazet@...il.com>
To:     Cong Wang <xiyou.wangcong@...il.com>, netdev@...r.kernel.org
Cc:     Jiri Pirko <jiri@...lanox.com>, Jamal Hadi Salim <jhs@...atatu.com>
Subject: Re: [Patch net-next] act_mirred: get rid of mirred_list_lock
 spinlock

On Thu, 2017-11-30 at 14:53 -0800, Cong Wang wrote:
> TC actions are no longer freed in RCU callbacks and we should
> always have RTNL lock, so this spinlock is no longer needed.
> 
> Cc: Jiri Pirko <jiri@...lanox.com>
> Cc: Jamal Hadi Salim <jhs@...atatu.com>
> Signed-off-by: Cong Wang <xiyou.wangcong@...il.com>
> ---
>  net/sched/act_mirred.c | 8 --------
>  1 file changed, 8 deletions(-)
> 
> diff --git a/net/sched/act_mirred.c b/net/sched/act_mirred.c
> index fe6489f9c3cf..2c51952bf2d4 100644
> --- a/net/sched/act_mirred.c
> +++ b/net/sched/act_mirred.c
> @@ -29,7 +29,6 @@
>  #include <net/tc_act/tc_mirred.h>
>  
>  static LIST_HEAD(mirred_list);
> -static DEFINE_SPINLOCK(mirred_list_lock);
>  
>  static bool tcf_mirred_is_act_redirect(int action)
>  {
> @@ -55,13 +54,10 @@ static void tcf_mirred_release(struct tc_action
> *a, int bind)
>  	struct tcf_mirred *m = to_mirred(a);
>  	struct net_device *dev;
>  
> -	/* We could be called either in a RCU callback or with RTNL
> lock held. */
> -	spin_lock_bh(&mirred_list_lock);
>  	list_del(&m->tcfm_list);
>  	dev = rcu_dereference_protected(m->tcfm_dev, 1);

If RTNL is held at this point, I suggest to use
rtnl_dereference() instead of rcu_dereference_protected() to get proper
lockdep coverage.


>  	if (dev)
>  		dev_put(dev);
> -	spin_unlock_bh(&mirred_list_lock);
>  }
>