Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
In-Reply-To: <20180830030420.GB16896@lunn.ch>
References: <20180830004046.9417-1-mdf@kernel.org> <20180830004046.9417-2-mdf@kernel.org>
 <20180830030420.GB16896@lunn.ch>
From:   Moritz Fischer <moritz.fischer@ettus.com>
Date:   Thu, 30 Aug 2018 10:21:03 -0700
Message-ID: <CAAtXAHe_fX9_XExvFj20a0L-=b0Z1gX410=wnR=P2DjO-hEByA@mail.gmail.com>
Subject: Re: [PATCH net-next 1/3] net: nixge: Add support for fixed-link subnodes
To:     Andrew Lunn <andrew@lunn.ch>
Cc:     "David S. Miller" <davem@davemloft.net>,
        Kees Cook <keescook@chromium.org>,
        Florian Fainelli <f.fainelli@gmail.com>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        netdev@vger.kernel.org, Alex Williams <alex.williams@ni.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Andrew,

On Wed, Aug 29, 2018 at 8:04 PM, Andrew Lunn <andrew@lunn.ch> wrote:
> On Wed, Aug 29, 2018 at 05:40:44PM -0700, Moritz Fischer wrote:
>> Add support for fixed-link cases where no MDIO is
>> actually required to run the device.
>> In that case no MDIO bus is instantiated since the
>> actual registers are not available in hardware.
>
> Hi Moritz
>
> There are a few different use cases here:
>
> The hardware is missing MDIO - You need fixed-link.

Agreed.
>
> The hardware has MDIO, but you don't have a PHY connected on it, and
> use fixed link.

Since it's an FPGA design in that case we'd probably build the hardware without
MDIO to save resources.

> The hardware has MDIO, and it is used e.g. for an Ethernet switch, or
> a PHY for another Ethernet interface. Plus you need fixed link.
We haven't had that yet but I can see that happen.
>
> The binding typically looks like:
>
> &fec1 {
>         phy-mode = "rmii";
>         pinctrl-names = "default";
>         pinctrl-0 = <&pinctrl_fec1>;
>         status = "okay";
>
>         fixed-link {
>                 speed = <100>;
>                 full-duplex;
>         };
>
>         mdio1: mdio {
>                 #address-cells = <1>;
>                 #size-cells = <0>;
>                 status = "okay";
>
>                 switch0: switch0@0 {
>                         compatible = "marvell,mv88e6085";
>                         pinctrl-names = "default";
>                         pinctrl-0 = <&pinctrl_switch>;
>                         reg = <0>;
>                         eeprom-length = <512>;
>                         interrupt-parent = <&gpio3>;
>
> It is important you have the mdio subnode, with PHYs and switches as
> children. The driver currently gets this wrong, it uses
> pdev->dev.of_node.

Oh, whoops. Yeah I should look into that. Any good examples of drivers doing
it right? Is the one going with the DT snippet above a good example?
>
> So the first patch should be to extend this behaviour. Look for a
> child node called mdio. If it exists, call nixge_mdio_setup() passing
> that child. Otherwise continue using pdev->dev.of_node, so you don't
> break backwards compatibility.

Ok will do.
>
> Then a patch adding support for fixed-link. If the mdio child node
> exists, you still need to register the MDIO bus. If there is no child
> node, but there is a fixed-link, skip registering the mdio bus with
> pdev->dev.of_node.
>
>         Andrew

Thanks for your feedback, much appreciated!

Moritz