Re: [PATCH] core/dev: set pkt_type after eth_type_trans() in dev_forward_skb()

[David Miller <davem@davemloft.net>]

From: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Date: Wed, 03 Jul 2013 16:53:52 +0200

> Le 03/07/2013 01:00, David Miller a écrit :
>> From: Isaku Yamahata <yamahata@valinux.co.jp>
>> Date: Tue,  2 Jul 2013 20:30:10 +0900
>>
>>> The dev_forward_skb() assignment of pkt_type should be done
>>> after the call to eth_type_trans().
>>>
>>> ip-encapsulated packets can be handled by localhost. But skb->pkt_type
>>> can be PACKET_OTHERHOST when packet comes via veth into ip tunnel
>>> device.
>>> In that case, the packet is dropped by ip_rcv().
>>> Although this example uses gretap. l2tp-eth also has same issue.
>>> For l2tp-eth case, add dummy device for ip address and ip l2tp
>>> command.
>>   ...
>>> Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
>>
>> Applied, but I had to adjust the patch because in net-next we use
>> a new helper function (skb_scrub_packet()) to clear things out from
>> dev_forward_skb().
> What about calling skb_scrub_packet() after eth_type_trans()?
> 
> I wonder if the same problem may happen the day gre will support
> x-netns,
> because skb_scrub_packet() is also called before eth_type_trans() in
> ip_tunnel_rcv().

That's a fine suggestion, I was just being overly conservative in my
forward-porting of his patch.