Re: [PATCH iproute2-next v3 0/2] f_flower: match on the number of vlan tags

[Stephen Hemminger <stephen@networkplumber.org>]

On Wed, 27 Apr 2022 17:32:00 +0300
Boris Sukholitko <boris.sukholitko@broadcom.com> wrote:

> Hi Stephen,
> 
> On Tue, Apr 26, 2022 at 08:11:42AM -0700, Stephen Hemminger wrote:
> > On Tue, 26 Apr 2022 12:14:15 +0300
> > Boris Sukholitko <boris.sukholitko@broadcom.com> wrote:
> >   
> > > Hi,
> > > 
> > > Our customers in the fiber telecom world have network configurations
> > > where they would like to control their traffic according to the number
> > > of tags appearing in the packet.
> > > 
> > > For example, TR247 GPON conformance test suite specification mostly
> > > talks about untagged, single, double tagged packets and gives lax
> > > guidelines on the vlan protocol vs. number of vlan tags.
> > > 
> > > This is different from the common IT networks where 802.1Q and 802.1ad
> > > protocols are usually describe single and double tagged packet. GPON
> > > configurations that we work with have arbitrary mix the above protocols
> > > and number of vlan tags in the packet.
> > > 
> > > The following patch series implement number of vlans flower filter. They
> > > add num_of_vlans flower filter as an alternative to vlan ethtype protocol
> > > matching. The end result is that the following command becomes possible:
> > > 
> > > tc filter add dev eth1 ingress flower \
> > >   num_of_vlans 1 vlan_prio 5 action drop
> > > 
> > > Also, from our logs, we have redirect rules such that:
> > > 
> > > tc filter add dev $GPON ingress flower num_of_vlans $N \
> > >      action mirred egress redirect dev $DEV
> > > 
> > > where N can range from 0 to 3 and $DEV is the function of $N.
> > > 
> > > Also there are rules setting skb mark based on the number of vlans:
> > > 
> > > tc filter add dev $GPON ingress flower num_of_vlans $N vlan_prio \
> > >     $P action skbedit mark $M
> > > 
> > > Thanks,
> > > Boris.
> > > 
> > > - v3: rebased to the latest iproute2-next
> > > - v2: add missing f_flower subject prefix
> > > 
> > > Boris Sukholitko (2):
> > >   f_flower: Add num of vlans parameter
> > >   f_flower: Check args with num_of_vlans
> > > 
> > >  tc/f_flower.c | 57 ++++++++++++++++++++++++++++++++++++---------------
> > >  1 file changed, 41 insertions(+), 16 deletions(-)  
> > 
> > Can you do this with BPF? instead of kernel change?  
> 
> You may have missed my reply to this question at:
> 
> https://lore.kernel.org/netdev/20220412104514.GB27480@noodle/
> 
> There is also Jamal's reply further at the thread:
> 
> https://lore.kernel.org/netdev/b2c83f63-a2e9-92a2-f262-3aae3491dfc3@mojatatu.com/
> 
> Thanks,
> Boris.

Thanks, there is a tradeoff here, if you add more logic to the kernel, it
impacts every user and creates long term technical debt. Your use case seemed
quite specific to single use case.

But also, it is an example of something where kernel already has the state
information and it might be hard to get with BPF.

Surprised that the people who do BPF at scale did not chime in on this discussion.