Re: [dpdk-dev] [PATCH v7 1/5] examples/l3fwd: fix LPM IPv6 subnets

["Walsh, Conor" <conor.walsh@intel.com>]

<snip>

> > > > The IPv6 subnets used were not within the 2001:200::/48 subnet
> > >
> > > {{32, 1, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}, 48, 0},
> > > {{32, 1, 2, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0}, 48, 1},
> > > etc...
> > >
> > > Err... all those subnets were exactly == 2001:200::/48.
> > > Is the issue that the example needs *separate* subnets taken from
> > > within 2001:200::/48 ?
> >
> > Hi David,
> > As it is ATM any IP with that range will match for all the rules for the 8 ports
> and the application cannot differentiate between them and it just returns
> the traffic to sender.
> > After this change the ports can be individually matched using the smaller
> /64 ranges for each port which are still within the given /48 subnet range
> (RFC5180).
> 
> Ok, thanks for confirming, could you update the commitlog to make it clear?

Hi David,

Thanks for your feedback on this patch and the other patches in the series, I will address them in v8.

Thanks,
Conor.

> 
> >
> > >
> > >
> > > > Changed to 2001:200:0:{0-7}::/64 where 0-7 is the port ID
> > > >
> > > > Fixes: 37afe381bde4 ("examples/l3fwd: use reserved IP addresses")
> > >
> > > And this looks like a Cc: stable@dpdk.org candidate.
> >
> > I can send a v8 and CC stable if you think that’s needed.
> 
> Yes please.
> I was waiting to look at the rest of the series, I have some comments
> that will require a v8 (from my pov).
> 
> 
> --
> David Marchand