[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <74469237-cea3-1eb6-1bd6-8d7ee55a8af5@prevas.dk>
Date: Mon, 16 Oct 2017 14:48:23 +0200
From: Martin Hundebøll <mnhu@...vas.dk>
To: Andrew Lunn <andrew@...n.ch>
CC: "David S . Miller" <davem@...emloft.net>, <netdev@...r.kernel.org>,
Vivien Didelot <vivien.didelot@...oirfairelinux.com>,
Florian Fainelli <f.fainelli@...il.com>
Subject: Re: [PATCH net-next] dsa: slave: support phy devices on external MII
bus
Hi Andrew,
On 2017-10-16 14:40, Andrew Lunn wrote:
>> /* internal MII */
>> mdio {
>> switch0phy1@1 {
>> reg = <1>;
>> };
>> };
>>
>> /* external MII */
>> mdio1 {
>> switch0phy0: switch0phy0@0 {
>> reg = <0>;
>> };
>
> Hi Martin
>
> You are missing a compatible string here. The binding document says:
>
> - mdio? : Container of PHYs and devices on the external MDIO
> bus. The node must contains a compatible string of
> "marvell,mv88e6xxx-mdio-external"
>
> Andrew
>
Yeah, I have it in my full dts file (attached snippet), but decided to
limit the commit-message version to keep it short(er). Should I update
the commit message to avoid confusing others?
The issue is really that dsa_slave_phy_connect() always uses the the
mdio bus associated with struct dsa_switch, even when the phy-handle
refers to a phy from another mdio bus.
Or am I missing something ?
// Martin
Download attachment "armada-388-gp.dts" of type "audio/vnd.dts" (2311 bytes)
Powered by blists - more mailing lists