Re: [net-next PATCH RFC] net: dsa: qca8k: make learning configurable and keep off if standalone

[Christian Marangi]

On Wed, Jun 28, 2023 at 02:49:53AM +0200, Christian Marangi wrote:
> On Tue, Jun 27, 2023 at 12:49:16PM +0300, Vladimir Oltean wrote:
> > On Mon, Jun 26, 2023 at 07:59:41PM +0200, Christian Marangi wrote:
> > > > Hmm, so ping works but this doesn't? That's strange, because send_uc_ipv4()
> > > > also pings. Have you run with bash -x to see why it fails?
> > > > 
> > > 
> > > I just run with bash -x and I also mod the script to not delete the
> > > tcpdump. Limiting the script to only this test the dump is just 2 ICMPv6
> > > packet and no output from tcpdump aside from
> > > 
> > > tcpdump: listening on lan1, link-type EN10MB (Ethernet), snapshot length 65535 bytes
> > > 3 packets captured
> > > 5 packets received by filter
> > > 0 packets dropped by kernel
> > > 
> > > I feel like this is important so I think I should focus on understanding
> > > why this doesn't work? Any clue?
> > 
> > No clue. I'd put a "bash" instruction in send_uc_ipv4(), which would act
> > as a sort of break point for the script (opens an interactive sub-shell),
> > then run it again with bash -x, manually repeat the command that failed,
> > investigate why it failed and hit Ctrl-d when I'm done.
> 
> Hi, I wasted a day to only notice that the problem was in busybox not
> supporting 0.x value and that is what selfttest script use. Another
> thing to check. Also the confusing part of this is that we don't check
> if ping_do return error and the test just fails (while in reality the
> ping command was never executed)
>

Also sorry for the double email... I forgot to add that it seems I need
to add a sleep 1 right after 

	mc_route_prepare $h1
	mc_route_prepare $rcv_if_name

in local_termination.sh. Think the switch needs some time to enable all
the port. Without the sleep 1 the first test just fails.

> Anyway I'm fixing all kind of bugs and I even found an even hw
> limitation with the FDB table with mgmt packet...
> 
> Also I implemented fdb_insolation following the implementation done on
> felix with reserving VID at the end...
> About this I wonder if it might be a good idea to expose some generic
> API from DSA?
> 
> qca8k require a reserved VID for VLAN unaware port and with
> fdb_isolation even more VID are reserved. DSA have no idea about this so
> I wonder if there is a chance of VID clash? I feel like we need
> something to declare a range of reserved VID so that they gets rejected
> when applied. (about this I think I should return -EINVAL if fdb/mbd
> insert are asked with a reserved VID)
> 
> Anyway by fixing that interval problem (enabling support in busybox as
> it's disabled by default in a OpenWRT system)
> This is the new output of the local_termination.sh 
> 
> TAP version 13
> 1..1
> # selftests: drivers/net/dsa: local_termination.sh
> # TEST: lan1: Unicast IPv4 to primary MAC address                     [ OK ]
> # TEST: lan1: Unicast IPv4 to macvlan MAC address                     [ OK ]
> # TEST: lan1: Unicast IPv4 to unknown MAC address                     [FAIL]
> #       reception succeeded, but should have failed
> # TEST: lan1: Unicast IPv4 to unknown MAC address, promisc            [ OK ]
> # TEST: lan1: Unicast IPv4 to unknown MAC address, allmulti           [FAIL]
> #       reception succeeded, but should have failed
> # TEST: lan1: Multicast IPv4 to joined group                          [ OK ]
> # TEST: lan1: Multicast IPv4 to unknown group                         [FAIL]
> #       reception succeeded, but should have failed
> # TEST: lan1: Multicast IPv4 to unknown group, promisc                [ OK ]
> # TEST: lan1: Multicast IPv4 to unknown group, allmulti               [ OK ]
> # TEST: lan1: Multicast IPv6 to joined group                          [ OK ]
> # TEST: lan1: Multicast IPv6 to unknown group                         [FAIL]
> #       reception succeeded, but should have failed
> # TEST: lan1: Multicast IPv6 to unknown group, promisc                [ OK ]
> # TEST: lan1: Multicast IPv6 to unknown group, allmulti               [ OK ]
> # TEST: br0: Unicast IPv4 to primary MAC address                      [ OK ]
> # TEST: br0: Unicast IPv4 to macvlan MAC address                      [ OK ]
> # TEST: br0: Unicast IPv4 to unknown MAC address                      [FAIL]
> #       reception succeeded, but should have failed
> # TEST: br0: Unicast IPv4 to unknown MAC address, promisc             [ OK ]
> # TEST: br0: Unicast IPv4 to unknown MAC address, allmulti            [FAIL]
> #       reception succeeded, but should have failed
> # TEST: br0: Multicast IPv4 to joined group                           [ OK ]
> # TEST: br0: Multicast IPv4 to unknown group                          [FAIL]
> #       reception succeeded, but should have failed
> # TEST: br0: Multicast IPv4 to unknown group, promisc                 [ OK ]
> # TEST: br0: Multicast IPv4 to unknown group, allmulti                [ OK ]
> # TEST: br0: Multicast IPv6 to joined group                           [ OK ]
> # TEST: br0: Multicast IPv6 to unknown group                          [FAIL]
> #       reception succeeded, but should have failed
> # TEST: br0: Multicast IPv6 to unknown group, promisc                 [ OK ]
> # TEST: br0: Multicast IPv6 to unknown group, allmulti                [ OK ]
> not ok 1 selftests: drivers/net/dsa: local_termination.sh # exit=1
> 
> Seems good aside from the reception secceeded that I think the kernel
> just drops right?
> 
> The switch have a way to control FLOOD per port but maybe the correct
> feature to use is the VLAN leaky? Where the setting can be set by both
> FDB/MBD and per port.
> 
> Sorry if I described some fix and implementation without proposing the
> patch but I would like some comments on what the tool returned.
> 
> -- 
> 	Ansuel

-- 
	Ansuel