On January 26, 2016 5:57:55 PM PST, Andrew Lunn <and...@lunn.ch> wrote:
>> Well, we do now, since that is one of the results used by
>> of_mdiobus_child_is_phy(),
>
>It uses it, but do not rely on it, since for backwards compatibility,
>we cannot assume it is there.
>
>You can never change an optional parameter to a mandatory parameter in
>DT. To do so breaks backwards compatibility.
>
>> What I would prefer seeing though is not removing nodes that have at
>> least two compatible strings, including one that is
>> "ethernet-phy-ieee802.3-c22", but those which have only one, like the
>> marvell ones that you patch, should have either an additional
>> "ethernet-phy-ieee802.3-c22", or none.
>
>So you are saying, if there is an "ethernet-phy-ieee802.3-c22" or an
>"ethernet-phy-ieee802.3-45" you can also have any other random junk,
>which we are going to ignore, since we have no way to verify, hence we
>have to assume it is broken, yet we need to be backwards compatible
>with it.
>
>Did you notice:
>
>+ { .compatible = "marvell,88e1310", },
>+ { .compatible = "marvell,88E1510", },
>
>No consistency with the 'e'. We would just be encouraging people to
>add more inconsistent stuff.
You are right, I am convinced now. If somebody needs to know the exact PHY ID a
priori, then you are also likely to use the more descriptive compatible string
including that ID.
So:
Acked-by: Florian Fainelli <f.faine...@gmail.com>
--
Florian
