[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <83403bcf-b6d8-f763-9269-682979a5619f@gmail.com>
Date: Fri, 18 Sep 2020 17:23:15 -0700
From: Florian Fainelli <f.fainelli@...il.com>
To: David Miller <davem@...emloft.net>, hongbo.wang@....com
Cc: xiaoliang.yang_1@....com, po.liu@....com, mingkai.hu@....com,
allan.nielsen@...rochip.com, claudiu.manoil@....com,
alexandru.marginean@....com, vladimir.oltean@....com,
leoyang.li@....com, andrew@...n.ch, vivien.didelot@...il.com,
jiri@...nulli.us, idosch@...sch.org, kuba@...nel.org,
vinicius.gomes@...el.com, nikolay@...ulusnetworks.com,
roopa@...ulusnetworks.com, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org, horatiu.vultur@...rochip.com,
alexandre.belloni@...tlin.com, UNGLinuxDriver@...rochip.com,
ivecera@...hat.com
Subject: Re: [PATCH v6 0/3] Add 802.1AD protocol support for dsa switch and
ocelot driver
On 9/18/2020 5:20 PM, David Miller wrote:
> From: hongbo.wang@....com
> Date: Wed, 16 Sep 2020 17:48:42 +0800
>
>> 1. Overview
>> a) 0001* is for support to set dsa slave into 802.1AD(QinQ) mode.
>> b) 0002* is for vlan_proto support for br_switchdev_port_vlan_add and br_switchdev_port_vlan_del.
>> c) 0003* is for setting QinQ related registers in ocelot switch driver, after applying this patch, the switch(VSC99599)'s port can enable or disable QinQ mode.
>
> You're going to have to update every single SWITCHDEV_PORT_ADD_OBJ handler
> and subsequent helpers to check the validate the protocol value.
>
> You also are going to have to make sure that every instantiated
> switchdev_obj_port_vlan object initializes the vlan protocol field
> properly.
>
> Basically, now that this structure has a new member, everything that
> operates on that object must be updated to handle the new protocol
> value.
>
> And I do mean everything.
>
> You can't just add the protocol handling to the locations you care
> about for bridging and DSA.
>
> You also have to more fully address the feedback given by Vladimir
> in patch #3. Are the expectations on the user side a Linux based
> expectation, or one specific about how this ASIC is expected to
> behave by default. It is very unclear what you are talking about
> when you say customer and ISP etc.
Switch vendors like to refer to the outer tag as ISP VLAN tag and inner
tag as customer VLAN tag, sometimes S-TAG and C-TAG terms are used, too...
--
Florian
Powered by blists - more mailing lists