[ovs-dev] [PATCH, v2] ovn: Allow SNAT traffic destined to router ip

[Chandra Sekhar Vejendla]

Ryan Moats/Omaha/IBM wrote on 07/18/2016 10:03:35 AM:

> From: Ryan Moats/Omaha/IBM
> To: Chandra Sekhar Vejendla/San Jose/IBM at IBMUS
> Cc: Guru Shetty <guru at ovn.org>, ovs dev <dev at openvswitch.org>
> Date: 07/18/2016 10:03 AM
> Subject: Re: [ovs-dev] [PATCH, v2] ovn: Allow SNAT traffic destined
> to router ip
>
> "dev" <dev-bounces at openvswitch.org> wrote on 07/18/2016 11:20:59 AM:
>
> > From: Chandra Sekhar Vejendla/San Jose/IBM at IBMUS
> > To: Guru Shetty <guru at ovn.org>
> > Cc: ovs dev <dev at openvswitch.org>
> > Date: 07/18/2016 11:21 AM
> > Subject: Re: [ovs-dev] [PATCH, v2] ovn: Allow SNAT traffic destined
> > to router ip
> > Sent by: "dev" <dev-bounces at openvswitch.org>
> >
> > Guru Shetty <guru at ovn.org> wrote on 07/18/2016 08:03:52 AM:
> >
> > > From: Guru Shetty <guru at ovn.org>
> > > To: Chandra Sekhar Vejendla/San Jose/IBM at IBMUS
> > > Cc: ovs dev <dev at openvswitch.org>
> > > Date: 07/18/2016 08:04 AM
> > > Subject: Re: [ovs-dev] [PATCH, v2] ovn: Allow SNAT traffic destined
> > > to router ip
> > >
> > > On 18 July 2016 at 05:22, Chandra S Vejendla <csvejend at us.ibm.com>
wrote:
> > > When router ip is used as SNAT IP, traffic destined to router
> > > ip should not be dropped
> > >
> > > Thank you for the fix. You will need to add your Signed-off-by. Can
> > > you also add a "Fixes:" tag in commit message. Since this is a
> > > regression, I wonder whether we should add a simple unit test that
> > > looks at the generated logical flows to make sure that there is no
> > > "drop" for the SNAT IP address.
> > >
> >
> > I'll send out patch v3 with 'Signed-off-by' & 'Fixes' tag.
> > Yes, i agree that we should have a unit test for this. If its fine with
you
> > I would like to cover the unit test case in a different patch.
>
> I'm sorry, but *I* want to see the test in the same patch.  We've
> already had cases where regressions were created because of
> lacking tests, and I'd like to avoid it, especially in cases like this
> where we are fixing a regression that could happen again.
>
> Ryan

Fair enough. I'll send out patch v4 with the test