Re: [patch net-next-2.6 2/8] bonding: register slave pointer for rx_handler

["Nicolas de Pesloüan" <nicolas.2p.debian@gmail.com>]

Le 05/03/2011 11:29, Jiri Pirko a écrit :
> Register slave pointer as rx_handler data. That would eventually prevent
> need to loop over slave devices to find the right slave.
>
> Use synchronize_net to ensure that bond_handle_frame does not get slave
> structure freed when working with that.
>
> Signed-off-by: Jiri Pirko<jpirko@redhat.com>
> ---
>   drivers/net/bonding/bond_main.c |   17 +++++++++++------
>   drivers/net/bonding/bonding.h   |    3 +++
>   2 files changed, 14 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c
> index 0592e6d..1c19368 100644
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -1495,21 +1495,22 @@ static bool bond_should_deliver_exact_match(struct sk_buff *skb,
>
>   static struct sk_buff *bond_handle_frame(struct sk_buff *skb)
>   {
> -	struct net_device *slave_dev;
> +	struct slave *slave;
>   	struct net_device *bond_dev;
>
>   	skb = skb_share_check(skb, GFP_ATOMIC);
>   	if (unlikely(!skb))
>   		return NULL;
> -	slave_dev = skb->dev;
> -	bond_dev = ACCESS_ONCE(slave_dev->master);
> +
> +	slave = bond_slave_get_rcu(skb->dev);
> +	bond_dev = ACCESS_ONCE(slave->dev->master);
>   	if (unlikely(!bond_dev))
>   		return skb;
>
>   	if (bond_dev->priv_flags&  IFF_MASTER_ARPMON)
> -		slave_dev->last_rx = jiffies;
> +		slave->dev->last_rx = jiffies;
>
> -	if (bond_should_deliver_exact_match(skb, slave_dev, bond_dev)) {
> +	if (bond_should_deliver_exact_match(skb, slave->dev, bond_dev)) {
>   		skb->deliver_no_wcard = 1;
>   		return skb;
>   	}

Up to this point, it looks like cleanup, and unrelated to the title and description of the patch.

Anyway, the cleanup looks good to me. Should just be in a separate patch.

> @@ -1703,7 +1704,8 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev)
>   		pr_debug("Error %d calling netdev_set_bond_master\n", res);
>   		goto err_restore_mac;
>   	}
> -	res = netdev_rx_handler_register(slave_dev, bond_handle_frame, NULL);
> +	res = netdev_rx_handler_register(slave_dev, bond_handle_frame,
> +					 new_slave);

And using rx_handler data for this purpose sounds good to me.

Reviewed-by: Nicolas de Pesloüan <nicolas.2p.debian@free.fr>

>   	if (res) {
>   		pr_debug("Error %d calling netdev_rx_handler_register\n", res);
>   		goto err_unset_master;
> @@ -1925,6 +1927,7 @@ err_close:
>
>   err_unreg_rxhandler:
>   	netdev_rx_handler_unregister(slave_dev);
> +	synchronize_net();
>
>   err_unset_master:
>   	netdev_set_bond_master(slave_dev, NULL);
> @@ -2108,6 +2111,7 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev)
>   	}
>
>   	netdev_rx_handler_unregister(slave_dev);
> +	synchronize_net();
>   	netdev_set_bond_master(slave_dev, NULL);
>
>   	slave_disable_netpoll(slave);
> @@ -2222,6 +2226,7 @@ static int bond_release_all(struct net_device *bond_dev)
>   		}
>
>   		netdev_rx_handler_unregister(slave_dev);
> +		synchronize_net();
>   		netdev_set_bond_master(slave_dev, NULL);
>
>   		slave_disable_netpoll(slave);
> diff --git a/drivers/net/bonding/bonding.h b/drivers/net/bonding/bonding.h
> index ff4e269..1aac5cd 100644
> --- a/drivers/net/bonding/bonding.h
> +++ b/drivers/net/bonding/bonding.h
> @@ -264,6 +264,9 @@ struct bonding {
>   #endif /* CONFIG_DEBUG_FS */
>   };
>
> +#define bond_slave_get_rcu(dev) \
> +	((struct slave *) rcu_dereference(dev->rx_handler_data))
> +
>   /**
>    * Returns NULL if the net_device does not belong to any of the bond's slaves
>    *