Re: [PATCH net] net: dsa: mv88e6xxx: Verify after ATU Load ops

[Andrew Lunn]

On Wed, Mar 05, 2025 at 12:44:54PM -0500, Joseph Huang wrote:
> On 3/5/2025 10:14 AM, Andrew Lunn wrote:
> > On Tue, Mar 04, 2025 at 06:53:51PM -0500, Joseph Huang wrote:
> > > ATU Load operations could fail silently if there's not enough space
> > > on the device to hold the new entry.
> > > 
> > > Do a Read-After-Write verification after each fdb/mdb add operation
> > > to make sure that the operation was really successful, and return
> > > -ENOSPC otherwise.
> > 
> > Please could you add a description of what the user sees when the ATU
> > is full. What makes this a bug which needs fixing? I would of thought
> > at least for unicast addresses, the switch has no entry for the
> > destination, so sends the packet to the CPU. The CPU will then
> > software bridge it out the correct port. Reporting ENOSPC will not
> > change that.
> 
> Hi Andrew,
> 
> What the user will see when the ATU table is full depends on the unknown
> flood setting. If a user has unknown multicast flood disabled, what the user
> will see is that multicast packets are dropped when the ATU table is full.
> In other words, IGMP snooping is broken when the ATU Load operation fails
> silently.

Please add this to the commit message. This describes the real problem
being fixed, which is what somebody reading the commit message wants
to know.

   Andrew