[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <583cd9b6-06e1-4ed7-8a24-c977b2001e20@lunn.ch>
Date: Tue, 5 Sep 2023 14:23:29 +0200
From: Andrew Lunn <andrew@...n.ch>
To: Vladimir Oltean <olteanv@...il.com>
Cc: Lukasz Majewski <lukma@...x.de>, Eric Dumazet <edumazet@...gle.com>,
davem@...emloft.net, Paolo Abeni <pabeni@...hat.com>,
Woojung Huh <woojung.huh@...rochip.com>, Tristram.Ha@...rochip.com,
Florian Fainelli <f.fainelli@...il.com>,
Jakub Kicinski <kuba@...nel.org>, UNGLinuxDriver@...rochip.com,
George McCollister <george.mccollister@...il.com>,
Oleksij Rempel <o.rempel@...gutronix.de>, netdev@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 RFC 4/4] net: dsa: hsr: Provide generic HSR
ksz_hsr_{join|leave} functions
> > And then we would have one port with SW HSR and another one with HW
> > HSR?
>
> No. One HSR device (hsr0, with 2 member ports) with offload and one
> HSR device (hsr1, with 2 member ports) without offload (see (b) below).
I just wanted to comment that offloading is about taking what Linux
can do in software and getting the hardware to do it. Linux should
happily allow two HSR devices, working in software. If you can only
offload one of them, Linux should continue to do the other one in
software.
So please do follow what Vladimir is suggesting.
Andrew
Powered by blists - more mailing lists