Re: [PATCH net-next] net: introduce per netns packet chains

[Sabrina Dubroca <sd@queasysnail.net>]

The patch mostly looks ok to me, except:


2025-03-17, 17:37:42 +0100, Paolo Abeni wrote:
>  /**
> - * dev_nit_active - return true if any network interface taps are in use
> + * dev_nit_active_rcu - return true if any network interface taps are in use
> + *
> + * The caller must held the RCU lock

typo: s/held/hold/


>   *
>   * @dev: network device to check for the presence of taps
>   */
> -bool dev_nit_active(struct net_device *dev)
> +bool dev_nit_active_rcu(struct net_device *dev)
>  {
> -	return !list_empty(&net_hotdata.ptype_all) ||
> +	return !list_empty(&dev_net_rcu(dev)->ptype_all) ||
>  	       !list_empty(&dev->ptype_all);
>  }
> -EXPORT_SYMBOL_GPL(dev_nit_active);
> +EXPORT_SYMBOL_GPL(dev_nit_active_rcu);
>  
>  /*
>   *	Support routine. Sends outgoing frames to any network
> @@ -2481,11 +2497,10 @@ EXPORT_SYMBOL_GPL(dev_nit_active);
>  
>  void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev)
>  {
> -	struct list_head *ptype_list = &net_hotdata.ptype_all;
> +	struct list_head *ptype_list = &dev_net_rcu(dev)->ptype_all;
>  	struct packet_type *ptype, *pt_prev = NULL;
>  	struct sk_buff *skb2 = NULL;
>  
> -	rcu_read_lock();

I can't convince myself that all callers of dev_queue_xmit_nit are
under RCU, specifically coming from drivers/net/wan/hdlc_x25.c
(x25_data_transmit). Maybe keep this rcu_read_lock here?

-- 
Sabrina