Re: [PATCH RFC net-next 0/3] Multi-CPU DSA support

On Sat, 24 Aug 2019 18:44:44 +0300
Vladimir Oltean [off-list ref] wrote:

Just to be clear. You can argue that such switches are weird, and
that's ok. Just want to understand the general type of hardware for
which such a patch is intended.

Vladimir,

the general part should solve for devices like Turris 1.x (qca8k) and
Turris Omnia (mv88e6xxx). In these devices the switch is connected to
CPU via 2 ports, and 5 ports are connected to RJ-45s.

I answered Andrew's question about the receive path in previous mail.
To your other question I still would have to think about, but the
general idea is that for other types of frames the switch driver
should only use one CPU port, so that no frame would reach CPU 2 times.

I shall send proposed implementation for mv88e6xxx in next version,
perhaps this night.

Marek