On Wed, Mar 10, 2021 at 01:30:33AM +0100, Andrew Lunn wrote: > On Tue, Mar 09, 2021 at 10:28:11PM +0100, Tobias Waldekranz wrote: > > Hi Tobias > > As with Florian, i've not been following the discussion.
I'm afraid there isn't much context for these patches, except Tobias pointing out that some of the command sequences I've since spelled out in the selftest produce results that should be refused but aren't. But this discussion has actually branched off and into the weeds from a completely different problem. Tobias reported that DSA attempts to install VLANs for 8021q uppers into the VTU when the mv88e6xxx ports are standalone, and the driver doesn't like that. This isn't due to an invalid constellation, but instead due to improper management of a valid one. I've attempted to solve those issues through these three patches, which at the moment are not properly grouped together, but I think I will pull them out and send them separately, and let Tobias finish the central rejection of invalid VLAN constellations: https://patchwork.kernel.org/project/netdevbpf/patch/20210308135509.3040286-1-olte...@gmail.com/ https://patchwork.kernel.org/project/netdevbpf/patch/20210309021657.3639745-4-olte...@gmail.com/ https://patchwork.kernel.org/project/netdevbpf/patch/20210309021657.3639745-5-olte...@gmail.com/ It would be nice if you could take a look at those three patches before I resend them tomorrow, also for Tobias to let me know if they fix the issue he initially reported.
