It looks like branch-2.1 didn't get the following commit that did make
it into branch-2.2. It may fix the problem. Can you try
cherry-picking it? If that solves it then I'm willing to backport it.
commit b73c85181df9cc38231a42d6f8095dcb604d230a
Author: Simon Horman <ho...@verge.net.au>
Date: Wed Jan 15 17:17:02 2014 +0900
netdev-linux: Read packet auxdata to obtain vlan_tid
If VLAN acceleration is used when the kernel receives a packet
then the outer-most VLAN tag will not be present in the packet
when it is received by netdev-linux. Rather, it will be present
in auxdata.
This patch uses recvmsg() instead of recv() to read auxdata for
each packet and if the vlan_tid is set then it is added to the packet.
Adding the vlan_tid makes use of headroom available
in the buffer parameter of rx_recv.
Signed-off-by: Simon Horman <ho...@verge.net.au>
Co-authored-by: Ben Pfaff <b...@nicira.com>
Signed-off-by: Ben Pfaff <b...@nicira.com>
On Fri, Jun 27, 2014 at 01:01:29PM -0400, Bob Cyphers wrote:
> I'm running
>
> Ubuntu 12.04.4 with Linux 3.5.0-51-generic
>
> Open vSwitch 2.1.2
>
> Intel niantic driver ixgbe 3.21.2
>
>
>
>
>
> On Fri, Jun 27, 2014 at 12:54 PM, Ben Pfaff <b...@nicira.com> wrote:
>
> > On Thu, Jun 26, 2014 at 12:57:54PM -0400, Bob Cyphers wrote:
> > > I'm attempting to test OVS as a gateway between VLAN and MPLS networks,
> > > starting with forwarding frames from one physical port to another. What
> > I
> > > have observed is that I can't make a rule that matches on VLAN ID (or
> > > dl_type=0x8100) when running userspace datapath.
> > >
> > > Is this the expected behavior with userspace datapath?
> >
> > I seem to recall that some old Linux kernels didn't properly pass
> > VLAN-tagged packets through to userspace. What kernel version are you
> > using?
> >
_______________________________________________
discuss mailing list
discuss@openvswitch.org
http://openvswitch.org/mailman/listinfo/discuss
