Re: [PATCH v8 net-next 05/12] net: dsa: propagate the locked flag down through the DSA layer

From: netdev
Date: Thu Oct 20 2022 - 14:47:53 EST


On 2022-10-20 15:35, Vladimir Oltean wrote:
On Thu, Oct 20, 2022 at 04:24:16PM +0300, Ido Schimmel wrote:
On Thu, Oct 20, 2022 at 04:02:24PM +0300, Vladimir Oltean wrote:
> On Tue, Oct 18, 2022 at 06:56:12PM +0200, Hans J. Schultz wrote:
> > @@ -3315,6 +3316,7 @@ static int dsa_slave_fdb_event(struct net_device *dev,
> > struct dsa_port *dp = dsa_slave_to_port(dev);
> > bool host_addr = fdb_info->is_local;
> > struct dsa_switch *ds = dp->ds;
> > + u16 fdb_flags = 0;
> >
> > if (ctx && ctx != dp)
> > return 0;
> > @@ -3361,6 +3363,9 @@ static int dsa_slave_fdb_event(struct net_device *dev,
> > orig_dev->name, fdb_info->addr, fdb_info->vid,
> > host_addr ? " as host address" : "");
> >
> > + if (fdb_info->locked)
> > + fdb_flags |= DSA_FDB_FLAG_LOCKED;
>
> This is the bridge->driver direction. In which of the changes up until
> now/through which mechanism will the bridge emit a
> SWITCHDEV_FDB_ADD_TO_DEVICE with fdb_info->locked = true?

I believe it can happen in the following call chain:

br_handle_frame_finish
br_fdb_update // p->flags & BR_PORT_MAB
fdb_notify
br_switchdev_fdb_notify

This can happen with Spectrum when a packet ingresses via a locked port
and incurs an FDB miss in hardware. The packet will be trapped and
injected to the Rx path where it should invoke the above call chain.

Ah, so this is the case which in mv88e6xxx would generate an ATU
violation interrupt; in the Spectrum case it generates a special packet.
Right now this packet isn't generated, right?

I think we have the same thing in ocelot, a port can be configured to
send "learn frames" to the CPU.

Should these packets be injected into the bridge RX path in the first
place? They reach the CPU because of an FDB miss, not because the CPU
was the intended destination.

Just to add to it, now that there is a u16 for flags in the bridge->driver
direction, making it easier to add such flags, I expect that for the
mv88e6xxx driver there shall be a 'IS_DYNAMIC' flag also, as authorized
hosts will have their authorized FDB entries added with dynamic entries...

Now as the bridge will not be able to refresh such authorized FDB entries
based on unicast incoming traffic on the locked port in the offloaded case,
besides we don't want the CPU to do such in this case anyway, to keep the
authorized line alive without having to reauthorize in like every 5 minutes,
the driver needs to do the ageing (and refreshing) of the dynamic entry
added from userspace. When the entry "ages" out, there is the HoldAt1
feature and Age Out Violations that should be used to tell userspace
(plus bridge) that this authorization has been removed by the driver as
the host has gone quiet.

So all in all, there is the need of another flag from
userspace->bridge->driver, telling that we want a dynamic ATU entry (with
mv88e6xxx it will start at age 7).