Re: [B.A.T.M.A.N.] [PATCH 2/4] batman-adv: Speed up dat by snooping received ip traffic

On Friday 12 February 2016 14:52:34 Andreas Pape wrote:

From cc88159dcf18f4b8310414d2d71635fad76bf5bb Mon Sep 17 00:00:00 2001 From: Andreas Pape apape@phoenixcontact.com Date: Fri, 12 Feb 2016 11:03:10 +0100 Subject: [PATCH 2/4] batman-adv: Speed up dat by snooping received ip traffic

This patch shall speed up dat by snooping all incoming ip traffic instead of only relying on ARP handling. This shall especially increase the probability that a gateway into a backbone network already has a fitting dat entry to answer incoming arp requests directly coming from the backbone network.

Signed-off-by: Andreas Pape apape@phoenixcontact.com

net/batman-adv/distributed-arp-table.c | 18 ++++++++++++++++++ net/batman-adv/distributed-arp-table.h | 8 +++++++- net/batman-adv/soft-interface.c | 21 ++++++++++++++++++++- 3 files changed, 45 insertions(+), 2 deletions(-)

diff --git a/net/batman-adv/distributed-arp-table.c b/net/batman-adv/distributed-arp-table.c index 93893bf..4e64e6c 100644 --- a/net/batman-adv/distributed-arp-table.c +++ b/net/batman-adv/distributed-arp-table.c @@ -362,6 +362,24 @@ out: batadv_dat_entry_put(dat_entry); }

+/**

• • batadv_dat_entry_check - check and update a dat entry
• • @bat_priv: the bat priv with all the soft interface information
• • @ip: ipv4 to add/edit
• • @mac_addr: mac address to assign to the given ipv4
• • @vid: VLAN identifier
• • checks additionally, if dat is enabled. can be called from other

modules.

• */

+void batadv_dat_entry_check(struct batadv_priv *bat_priv, __be32 ip,

•                            u8 *mac_addr, unsigned short vid)
  

+{

•   if(!atomic_read(&bat_priv->distributed_arp_table))
  

•           return;
  

•   batadv_dat_entry_add(bat_priv, ip, mac_addr, vid);
  

+}

#ifdef CONFIG_BATMAN_ADV_DEBUG

/** diff --git a/net/batman-adv/distributed-arp-table.h b/net/batman-adv/distributed-arp-table.h index 813ecea..a2ab16b 100644 --- a/net/batman-adv/distributed-arp-table.h +++ b/net/batman-adv/distributed-arp-table.h @@ -80,7 +80,8 @@ batadv_dat_init_own_addr(struct batadv_priv *bat_priv, int batadv_dat_init(struct batadv_priv *bat_priv); void batadv_dat_free(struct batadv_priv *bat_priv); int batadv_dat_cache_seq_print_text(struct seq_file *seq, void *offset);

+void batadv_dat_entry_check(struct batadv_priv *bat_priv, __be32 ip,

•                            u8 *mac_addr, unsigned short vid);
  

/**

• batadv_dat_inc_counter - increment the correct DAT packet counter
• @bat_priv: the bat priv with all the soft interface information

@@ -173,6 +174,11 @@ static inline void batadv_dat_inc_counter(struct batadv_priv *bat_priv, { }

+void batadv_dat_entry_check(struct batadv_priv *bat_priv, __be32 ip,

•                            u8 *mac_addr, unsigned short vid)
  

+{ +}

#endif /* CONFIG_BATMAN_ADV_DAT */

#endif /* _NET_BATMAN_ADV_DISTRIBUTED_ARP_TABLE_H_ */ diff --git a/net/batman-adv/soft-interface.c b/net/batman-adv/soft-interface.c index 0710379..41d7987 100644 --- a/net/batman-adv/soft-interface.c +++ b/net/batman-adv/soft-interface.c @@ -28,6 +28,7 @@ #include <linux/fs.h> #include <linux/if_ether.h> #include <linux/if_vlan.h> +#include <linux/ip.h> #include <linux/jiffies.h> #include <linux/kernel.h> #include <linux/kref.h> @@ -390,6 +391,7 @@ void batadv_interface_rx(struct net_device *soft_iface, __be16 ethertype = htons(ETH_P_BATMAN); struct vlan_ethhdr *vhdr; struct ethhdr *ethhdr;

•   struct iphdr *iphdr;
    unsigned short vid;
    bool is_bcast;
  
  

@@ -412,11 +414,28 @@ void batadv_interface_rx(struct net_device *soft_iface, ethhdr = eth_hdr(skb);

    switch (ntohs(ethhdr->h_proto)) {

•   case ETH_P_IP:
  

•           iphdr = (struct iphdr *)(skb->data + ETH_HLEN);
  

•           /* snoop incoming traffic for dat update using the source
  

mac

•            * and source ip to speed up dat.
  

•            * Question: does this break the fundamental idea of
  

dat????

•            */
  

That is a really good question, although it doesn't belong in the code ;)

@Antonio, CC'ing you since this is more a design question/proposal and you may have thought about this yet.

Basically, doing this change means that we will put a lot of IP addresses in our cache which are not in our local network - typically, all Internet IP addresses along with the gateway backbone. Also these addresses will never be requested by ARP and are therefore practically just littering our cache. They are purged after 5 minutes so the impact may be reasonable, but still ...

Maybe there is a way to limit the entries to local networks? Also (and in general), should we have an upper limit how many entries we store in DAT? After applying this patch, doing a subnet ping scan can deplete the RAM in small routers I'm afraid. :)

(even now, that would be possible with fake ARP replies I guess)

Also, why don't you check the ip destination as well while at it?

Cheers, Simon