Re: [PATCH net v2 2/7] net: dsa: mt7530: fix trapping frames with multiple CPU ports on MT7530

[Vladimir Oltean]

On Tue, Jun 13, 2023 at 08:18:58PM +0300, Vladimir Oltean wrote:
> On Tue, Jun 13, 2023 at 08:14:35PM +0300, Arınç ÜNAL wrote:
> > Actually, having only "net: dsa: introduce preferred_default_local_cpu_port
> > and use on MT7530" backported is an enough solution for the current stable
> > kernels.
> > 
> > When multiple CPU ports are defined on the devicetree, the CPU_PORT bits
> > will be set to port 6. The active CPU port will also be port 6.
> > 
> > This would only become an issue with the changing the DSA conduit support.
> > But that's never going to happen as this patch will always be on the kernels
> > that support changing the DSA conduit.
> 
> Aha, ok. I thought that device trees with CPU port 5 exclusively defined
> also exist in the wild. If not, and this patch fixes a theoretical only
> issue, then it is net-next material.

On second thought, compatibility with future device trees is the reason
for this patch set, so that should equally be a reason for keeping this
patch in a "net" series.

If I understand you correctly, port 5 should have worked since commit
c8b8a3c601f2 ("net: dsa: mt7530: permit port 5 to work without port 6 on
MT7621 SoC"), and it did, except for trapping, right?

So how about settling on that as a more modest Fixes: tag, and
explaining clearly in the commit message what's affected?