Re: [B.A.T.M.A.N.] [PATCH 2/3] batman-adv: introduce a boolean switch to enable/disable DAT

[Marek Lindner <lindner_marek@yahoo.de>]

On Tuesday, April 17, 2012 12:24:53 Antonio Quartulli wrote:
> +static void arp_change_timeout(struct net_device *soft_iface, bool
> enable_dat) +{
> +       struct in_device *in_dev = in_dev_get(soft_iface);
> +       if (!in_dev) {
> +               pr_err("Unable to set ARP parameters for the batman
> interface '%s'\n", +                      soft_iface->name);
> +               return;
> +       }
> +
> +       if (enable_dat) {
> +               /* Introduce a delay in the ARP state-machine transactions.
> +                * Entries will be kept in the ARP table for the default
> time +                * multiplied by 4
> +                */
> +               in_dev->arp_parms->base_reachable_time *=
> ARP_TIMEOUT_FACTOR; +               in_dev->arp_parms->gc_staletime *=
> ARP_TIMEOUT_FACTOR; +               in_dev->arp_parms->reachable_time *=
> ARP_TIMEOUT_FACTOR; +       } else {
> +               in_dev->arp_parms->base_reachable_time /=
> ARP_TIMEOUT_FACTOR; +               in_dev->arp_parms->gc_staletime /=
> ARP_TIMEOUT_FACTOR; +               in_dev->arp_parms->reachable_time /=
> ARP_TIMEOUT_FACTOR; +       }
> +
> +       in_dev_put(in_dev);
> +}
> +
> +void dat_sysfs_change(struct net_device *soft_iface)
> +{
> +       struct bat_priv *bat_priv = netdev_priv(soft_iface);
> +
> +       if (!bat_priv)
> +               return;
> +
> +       /* increase or decrease the ARP timeout */
> +       arp_change_timeout(soft_iface,
> +                          atomic_read(&bat_priv->distributed_arp_table));
> +}

AFAIK there is no check whether the received value actually changes something 
but arp_change_timeout() implicitely assumes so. I guess the following will 
break the timeouts:

echo 0 > /sys/class/net/bat0/mesh/distributed_arp_table
echo 0 > /sys/class/net/bat0/mesh/distributed_arp_table

Regards,
Marek