[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <d4ab92c8-8b70-ce26-e47f-d5172c8fd4e7@ti.com>
Date: Thu, 3 Dec 2020 18:22:05 +0200
From: Grygorii Strashko <grygorii.strashko@...com>
To: Andrew Lunn <andrew@...n.ch>, Vignesh Raghavendra <vigneshr@...com>
CC: "David S . Miller" <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>,
Jonathan Corbet <corbet@....net>, Jiri Pirko <jiri@...dia.com>,
<netdev@...r.kernel.org>, <linux-doc@...r.kernel.org>,
<linux-kernel@...r.kernel.org>,
Linux ARM Mailing List <linux-arm-kernel@...ts.infradead.org>
Subject: Re: [PATCH 1/4] net: ti: am65-cpsw-nuss: Add devlink support
On 03/12/2020 16:18, Andrew Lunn wrote:
>> We don't want to enable HW based switch support unless explicitly
>> asked by user.
>
> This is the key point. Why? Does individual ports when passed through
> the switch not work properly? Does it add extra latency/jitter?
When switch mode is enabled the forwarding is enabled by default and can't be completely
disabled, while in multi port mode every port and switch tables (ALE) configured so no packet
leaking between ports is happen.
The multi port is the requirement for us to have as default mode no mater to what upper interface
ports are attached to LAG, LRE (HSR/PRP) or bridge.
Switching between modes required significant Port and ALE reconfiguration there for
technical decision made and implemented to use parameter for mode change (by using devlink).
It also allows to keep user interface similar to what was implements for previous generation
of TI CPSW (am3/4/5).
--
Best regards,
grygorii
Powered by blists - more mailing lists