[net-next,2/2] openvswitch: Fix skb->protocol for vlan frames.

> On Nov 24, 2016, at 8:10 AM, Jiri Benc <jbenc@redhat.com> wrote:
> 
> On Tue, 22 Nov 2016 20:09:34 -0800, Jarno Rajahalme wrote:
>> Do not set skb->protocol to be the ethertype of the L3 header, unless
>> the packet only has the L3 header.  For a non-hardware offloaded VLAN
>> frame skb->protocol needs to be one of the VLAN ethertypes.
>> 
>> Any VLAN offloading is undone on the OVS netlink interface.  Due to
>> this all VLAN packets sent to openvswitch module from userspace are
>> non-offloaded.
> 
> This is exactly why I wanted to always accelerate the vlan tag, the
> same way it is done in other parts of the networking stack: to prevent
> all those weird corner cases.
> 
> Looks to me this is the only real way forward.
> 

I’m not sure what you suggest here. Obviously the kernel ABI can not be changed as existing userspace code expects upcalled packets to be non-accelerated. Also, if userspace pushes vlan headers, the packet will actually have them.

> This patch is wrong, it would leave skb->protocol as ETH_P_TEB for L2
> frames received via ARPHRD_NONE interface.
> 

Would this incremental fix this:



  Jarno

On Mon, 28 Nov 2016 14:29:39 -0800, Jarno Rajahalme wrote:
> I’m not sure what you suggest here. Obviously the kernel ABI can not
> be changed as existing userspace code expects upcalled packets to be
> non-accelerated. Also, if userspace pushes vlan headers, the packet
> will actually have them.

The user space API needs to be preserved, of course. I'm talking about
what happens internally in the kernel.

See this patchset: https://www.spinics.net/lists/netdev/msg398827.html

> Would this incremental fix this:
> 
> diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
> index 9be9fda..37f1bb9 100644
> --- a/net/openvswitch/flow.c
> +++ b/net/openvswitch/flow.c
> @@ -354,6 +354,8 @@ static int parse_vlan(struct sk_buff *skb, struct
> sw_flow_key *key) res = parse_vlan_tag(skb, &key->eth.vlan);
>  		if (res <= 0)
>  			return res;
> +		if (skb->protocol == htons(ETH_P_TEB))
> +			skb->protocol = key->eth.vlan.tpid;
>  	}
>  
>  	/* Parse inner vlan tag. */

I'll look at this tomorrow. But it seems we're adding more and more
hacks instead of cleaning up the vlan handling.

 Jiri

> On Nov 28, 2016, at 2:42 PM, Jiri Benc <jbenc@redhat.com> wrote:
> 
> On Mon, 28 Nov 2016 14:29:39 -0800, Jarno Rajahalme wrote:
>> I’m not sure what you suggest here. Obviously the kernel ABI can not
>> be changed as existing userspace code expects upcalled packets to be
>> non-accelerated. Also, if userspace pushes vlan headers, the packet
>> will actually have them.
> 
> The user space API needs to be preserved, of course. I'm talking about
> what happens internally in the kernel.
> 
> See this patchset: https://www.spinics.net/lists/netdev/msg398827.html
> 

I did not try to apply this series yet, but given the recent L3 changes maybe it needs a rebase?

>> Would this incremental fix this:
>> 
>> diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
>> index 9be9fda..37f1bb9 100644
>> --- a/net/openvswitch/flow.c
>> +++ b/net/openvswitch/flow.c
>> @@ -354,6 +354,8 @@ static int parse_vlan(struct sk_buff *skb, struct
>> sw_flow_key *key) res = parse_vlan_tag(skb, &key->eth.vlan);
>> 		if (res <= 0)
>> 			return res;
>> +		if (skb->protocol == htons(ETH_P_TEB))
>> +			skb->protocol = key->eth.vlan.tpid;
>> 	}
>> 
>> 	/* Parse inner vlan tag. */
> 
> I'll look at this tomorrow. But it seems we're adding more and more
> hacks instead of cleaning up the vlan handling.
> 

Right, I just noticed that the incremental only handles the VLAN case.

I’ll post a v2 later today with a proper fix that solves the immediate issue. IMO this should be fixed independently of the VLAN handling series for which I have no informed opinion yet.

  Jarno

> Jiri