[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20211207231449.bk5mxg3z2o7mmtzg@skbuf>
Date: Wed, 8 Dec 2021 01:14:49 +0200
From: Vladimir Oltean <olteanv@...il.com>
To: Andrew Lunn <andrew@...n.ch>
Cc: Ansuel Smith <ansuelsmth@...il.com>,
Vivien Didelot <vivien.didelot@...il.com>,
Florian Fainelli <f.fainelli@...il.com>,
"David S. Miller" <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>, linux-kernel@...r.kernel.org,
netdev@...r.kernel.org
Subject: Re: [net-next RFC PATCH 0/6] Add support for qca8k mdio rw in
Ethernet packet
On Tue, Dec 07, 2021 at 11:54:07PM +0100, Andrew Lunn wrote:
> > I considered a simplified form like this, but I think the tagger private
> > data will still stay in dp->priv, only its ownership will change.
>
> Isn't dp a port structure. So there is one per port?
Yes, but dp->priv is a pointer. The thing it points to may not
necessarily be per port.
> This is where i think we need to separate shared state from tagger
> private data. Probably tagger private data is not per port. Shared
> state between the switch driver and the tagger maybe is per port?
I don't know whether there's such a big difference between
"shared state" vs "private data"? The dp->priv model is flexible enough
to support both. For example, in tag_sja1105, dp->priv is a struct
sja1105_port. All struct sja1105_port of a switch have a common struct
sja1105_tagger_data *data pointer. We could certainly set up the
tag_ops->connect(dst) function to allocate memory in this way.
Powered by blists - more mailing lists