Allow fragmentation of VLAN packets traversing a bridge.

br_nf_dev_queue_xmit only checks for ETH_P_IP packets for fragmenting but not
VLAN packets. This results in dropping of large VLAN packets. This can be
observed when connection tracking is enabled. Connection tracking re-assembles
fragmented packets, and these have to be re-fragmented when transmitting out.

Signed-off-by: Saikiran Madugula <hummerbliss@gmail.com>
---
 net/bridge/br_netfilter.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

Saikiran Madugula wrote:
> br_nf_dev_queue_xmit only checks for ETH_P_IP packets for fragmenting but not
> VLAN packets. This results in dropping of large VLAN packets. This can be
> observed when connection tracking is enabled. Connection tracking re-assembles
> fragmented packets, and these have to be re-fragmented when transmitting out.
> 
> Signed-off-by: Saikiran Madugula <hummerbliss@gmail.com>
> ---
>  net/bridge/br_netfilter.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c
> index 3953ac4..941f702 100644
> --- a/net/bridge/br_netfilter.c
> +++ b/net/bridge/br_netfilter.c
> @@ -790,7 +790,7 @@ static unsigned int br_nf_local_out(unsigned int hook, struct sk_buff *skb,
>  
>  static int br_nf_dev_queue_xmit(struct sk_buff *skb)
>  {
> -	if (skb->protocol == htons(ETH_P_IP) &&
> +	if ((skb->protocol == htons(ETH_P_IP) || IS_VLAN_IP(skb)) &&
>  	    skb->len > skb->dev->mtu &&
>  	    !skb_is_gso(skb))
>  		return ip_fragment(skb, br_dev_queue_push_xmit);

Please add an additional check for skb->nfct != NULL to make sure
that this only refragments packets defragmented by conntrack.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patrick McHardy wrote:
> Saikiran Madugula wrote:
>> br_nf_dev_queue_xmit only checks for ETH_P_IP packets for fragmenting
>> but not
>> VLAN packets. This results in dropping of large VLAN packets. This can be
>> observed when connection tracking is enabled. Connection tracking
>> re-assembles
>> fragmented packets, and these have to be re-fragmented when
>> transmitting out.
>>
>> Signed-off-by: Saikiran Madugula <hummerbliss@gmail.com>
>> ---
>>  net/bridge/br_netfilter.c |    2 +-
>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c
>> index 3953ac4..941f702 100644
>> --- a/net/bridge/br_netfilter.c
>> +++ b/net/bridge/br_netfilter.c
>> @@ -790,7 +790,7 @@ static unsigned int br_nf_local_out(unsigned int
>> hook, struct sk_buff *skb,
>>  
>>  static int br_nf_dev_queue_xmit(struct sk_buff *skb)
>>  {
>> -    if (skb->protocol == htons(ETH_P_IP) &&
>> +    if ((skb->protocol == htons(ETH_P_IP) || IS_VLAN_IP(skb)) &&
>>          skb->len > skb->dev->mtu &&
>>          !skb_is_gso(skb))
>>          return ip_fragment(skb, br_dev_queue_push_xmit);
> 
> Please add an additional check for skb->nfct != NULL to make sure
> that this only refragments packets defragmented by conntrack.

Thanks for the feedback, skb->nfct is present only if CONFIG_NF_CONNTRACK or
CONFIG_NF_CONNTRACK_MODULE is defined. Was wondering if the entire check before
ip_fragment is necessary if NF_CONNTRACK is not defined. If not, I will post
updated patch as per your suggestion.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Saikiran Madugula wrote:
> Patrick McHardy wrote:
>>> -    if (skb->protocol == htons(ETH_P_IP) &&
>>> +    if ((skb->protocol == htons(ETH_P_IP) || IS_VLAN_IP(skb)) &&
>>>          skb->len > skb->dev->mtu &&
>>>          !skb_is_gso(skb))
>>>          return ip_fragment(skb, br_dev_queue_push_xmit);
>> Please add an additional check for skb->nfct != NULL to make sure
>> that this only refragments packets defragmented by conntrack.
> 
> Thanks for the feedback, skb->nfct is present only if CONFIG_NF_CONNTRACK or
> CONFIG_NF_CONNTRACK_MODULE is defined. Was wondering if the entire check before
> ip_fragment is necessary if NF_CONNTRACK is not defined. If not, I will post
> updated patch as per your suggestion.

Good point. Yes, everything related to fragmenting is only needed
with NF_CONNTRACK, so an additional ifdef makes sense.

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html