Re: [PATCH v3 net-next 13/24] net: dsa: Allow drivers to filter packets they can decode source port from

From: Vladimir Oltean <olteanv@gmail.com>
To: Andrew Lunn <andrew@lunn.ch>
Cc: Florian Fainelli <f.fainelli@gmail.com>,
	vivien.didelot@gmail.com, davem@davemloft.net,
	netdev <netdev@vger.kernel.org>,
	linux-kernel@vger.kernel.org,
	Georg Waibel <georg.waibel@sensor-technik.de>
Subject: Re: [PATCH v3 net-next 13/24] net: dsa: Allow drivers to filter packets they can decode source port from
Date: Sat, 13 Apr 2019 18:48:37 +0300	[thread overview]
Message-ID: <CA+h21hqNMCzFKkdoeB0snbXY1E+BD3Rp24geuD_ZbNVPJf4kuQ@mail.gmail.com> (raw)
In-Reply-To: <20190413153959.GE17901@lunn.ch>

On Sat, 13 Apr 2019 at 18:40, Andrew Lunn <andrew@lunn.ch> wrote:
>
> On Sat, Apr 13, 2019 at 04:28:11AM +0300, Vladimir Oltean wrote:
> > Frames get processed by DSA and redirected to switch port net devices
> > based on the ETH_P_XDSA multiplexed packet_type handler found by the
> > network stack when calling eth_type_trans().
> >
> > The running assumption is that once the DSA .rcv function is called, DSA
> > is always able to decode the switch tag in order to change the skb->dev
> > from its master.
> >
> > However there are tagging protocols (such as the new DSA_TAG_PROTO_SJA1105,
> > user of DSA_TAG_PROTO_8021Q) where this assumption is not completely
> > true, since switch tagging piggybacks on the absence of a vlan_filtering
> > bridge. Moreover, management traffic (BPDU, PTP) for this switch doesn't
> > rely on switch tagging, but on a different mechanism. So it would make
> > sense to at least be able to terminate that.
>
> Hi Vladimir
>
> Let me see if i get this correct.
>
> If the filter fails to match, the frame is received on the master
> interface? So BPDUs and PTT packets are going to go to the master
> interface?
>
> How does the bridge get these BPDUs, and associated to the correct
> slave port?
>
> How does the PTP core code get these frames, and associated to the
> correct slave port?
>
> You say there is a different mechanism to do this. Maybe a later patch
> i've not yet looked at. But cannot this mechanism be built into the
> tagger? That is what the tagger is there for, to demultiplex a frame
> to the correct slave. The current code assume the needed information
> is in the header, but there is nothing to stop it looking deeper into
> the packet if needed. So far, we have been reluctant for a tagger to
> call into the DSA driver, but if need be, it could happen.
>
>      Andrew

Hi Andrew,

Yes it's explained in a further patch. See the checks for
sja1105_is_link_local() in patch 18/24.
There's also a table in the documentation patch (23/24).

Thanks,
-Vladimir