Message-ID: <r247bmekxv2de7owpoam6kkscel25ugnneebzwsrv3j7u3lud7@ppuwdzwl4zi5>
Date: Sun, 10 Dec 2023 13:48:12 +0000
From: Alvin Šipraga <ALSI@...g-olufsen.dk>
To: Vladimir Oltean <vladimir.oltean@....com>
CC: "netdev@...r.kernel.org" <netdev@...r.kernel.org>, "David S. Miller"
	<davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski
	<kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>, Andrew Lunn
	<andrew@...n.ch>, Florian Fainelli <f.fainelli@...il.com>, Luiz Angelo Daros
 de Luca <luizluca@...il.com>, Madhuri Sripada
	<madhuri.sripada@...rochip.com>, Marcin Wojtas <mw@...ihalf.com>, Linus
 Walleij <linus.walleij@...aro.org>, Tobias Waldekranz
	<tobias@...dekranz.com>, Arun Ramadoss <arun.ramadoss@...rochip.com>,
	"Russell King (Oracle)" <rmk+kernel@...linux.org.uk>, Jonathan Corbet
	<corbet@....net>
Subject: Re: [PATCH net 3/4] docs: net: dsa: update user MDIO bus
 documentation

On Fri, Dec 08, 2023 at 09:35:17PM +0200, Vladimir Oltean wrote:
> There are people who are trying to push the ds->user_mii_bus feature
> past its sell-by date. I think part of the problem is the fact that the
> documentation presents it as this great functionality.
> 
> Adapt it to 2023, where we have phy-handle to render it useless, at
> least with OF.
> 
> Signed-off-by: Vladimir Oltean <vladimir.oltean@....com>
> ---
>  Documentation/networking/dsa/dsa.rst | 36 ++++++++++++++++++++++------
>  1 file changed, 29 insertions(+), 7 deletions(-)
> 
> diff --git a/Documentation/networking/dsa/dsa.rst b/Documentation/networking/dsa/dsa.rst
> index 676c92136a0e..2cd91358421e 100644
> --- a/Documentation/networking/dsa/dsa.rst
> +++ b/Documentation/networking/dsa/dsa.rst
> @@ -397,19 +397,41 @@ perspective::
>  User MDIO bus
>  -------------
>  
> -In order to be able to read to/from a switch PHY built into it, DSA creates an
> -user MDIO bus which allows a specific switch driver to divert and intercept
> -MDIO reads/writes towards specific PHY addresses. In most MDIO-connected
> -switches, these functions would utilize direct or indirect PHY addressing mode
> -to return standard MII registers from the switch builtin PHYs, allowing the PHY
> -library and/or to return link status, link partner pages, auto-negotiation
> -results, etc.
> +The framework creates an MDIO bus for user ports (``ds->user_mii_bus``) when
> +both methods ``ds->ops->phy_read()`` and ``ds->ops->phy_write()`` are present.
> +However, this pointer may also be populated by the switch driver during the
> +``ds->ops->setup()`` method, with an MDIO bus managed by the driver.
> +
> +Its role is to permit user ports to connect to a PHY (usually internal) when
> +the more general ``phy-handle`` property is unavailable (either because the
> +MDIO bus is missing from the OF description, or because probing uses
> +``platform_data``).
> +
> +In most MDIO-connected switches, these functions would utilize direct or
> +indirect PHY addressing mode to return standard MII registers from the switch
> +builtin PHYs, allowing the PHY library and/or to return link status, link
> +partner pages, auto-negotiation results, etc.
>  
>  For Ethernet switches which have both external and internal MDIO buses, the
>  user MII bus can be utilized to mux/demux MDIO reads and writes towards either
>  internal or external MDIO devices this switch might be connected to: internal
>  PHYs, external PHYs, or even external switches.
>  
> +When using OF, the ``ds->user_mii_bus`` can be seen as a legacy feature, rather
> +than core functionality. Since 2014, the DSA OF bindings support the
> +``phy-handle`` property, which is a universal mechanism to reference a PHY,
> +be it internal or external.
> +
> +New switch drivers are encouraged to require the more universal ``phy-handle``
> +property even for user ports with internal PHYs. This allows device trees to
> +interoperate with simpler variants of the drivers such as those from U-Boot,
> +which do not have the (redundant) fallback logic for ``ds->user_mii_bus``.

Considering this policy, should we not emphasize that ds->user_mii_bus
and ds->ops->phy_{read,write}() ought to be left unpopulated by new
drivers, with the remark that if a driver wants to set up an MDIO bus,
it should store the corresponding struct mii_bus pointer in its own
driver private data? Just to make things crystal clear.

Regardless I think this is good!

Reviewed-by: Alvin Šipraga <alsi@...g-olufsen.dk>

> +
> +The only use case for ``ds->user_mii_bus`` in new drivers would be for probing
> +on non-OF through ``platform_data``. In the distant future where this will be
> +possible through software nodes, there will be no need for ``ds->user_mii_bus``
> +in new drivers at all.
> +
>  Data structures
>  ---------------
>  
> -- 
> 2.34.1
>

Powered by blists - more mailing lists