On 1/7/20 1:46 AM, Vladimir Oltean wrote:

So we don't typically have the external MDIO controller be part of the
memory map of an Ethernet port, it has its own region in the SoC. But
this is not an external MDIO controller, and it's a completely
separate hardware instance with its own memory region not shared with
anybody else.

OK, I do not think I managed to express myself correctly, by external
MDIO controller I meant a controller that is used to interface with
off-chip (and sometimes on-chip, too) MDIO devices. That is an on-chip
MDIO controller technically, pardon me for not expressing myself more
clearly here.

And I think your need for registering the slave MDIO bus in the
Starfighter 2 switch driver, on top of the same memory region as the
GENET's MDIO bus, is just to work around the fact that the switch
pseudo-PHY MDIO address is fixed at 30 on some chips, and you need to
intercept those transactions by wrapping the mii_bus->read and
mii_bus->write methods of the master, otherwise none of this would
have been needed.

You are conflating two things that are not related to the point being
discussed. The SF2 switch used on 7445 and 7278 has some wrapping around
the Roboswitch original IP. The roboswitch has an internal MDIO
controller which is managed through the register space within the
switch. We added a MDIO controller to interface with both our internal
Gigabit PHYs as well as external MDIO devices would such thing exist.
The binding show the MDIO node as part of "switch_top" but not being
part of "core":
Documentation/devicetree/bindings/net/brcm,bcm7445-switch-v4.0.txt.

The same IP used as the MDIO controller (those two 32-bit words with CFG
and CMD) are the same as those that are used on GENET, hence the desire
to re-use the same piece of driver.

If I understand correctly, couldn't you have just overridden the
mii_bus->read and mii_bus->write ops of the master MDIO bus, without
registering another one?

Yes maybe, I do not think this is particularly relevant to this
discussion though.

quoted

quoted

quoted

Your commit message does not provide a justification for why this
abstraction (mii_bus) was not suitable or considered here. Do you think
that could be changed?

I'm sorry, was the mii_bus abstraction really not considered here?
Based on the stuff exported in this patch, an mii_bus is exactly what
I'm registering in 9/9, no?

I meat in the commit message, there is no justification why this was not
considered or used, by asking you ended up providing one, that is
typically what one would expect to find to explain why something was/was
not considered. It's fine, the code is merge, I won't object or require
you to use a mii_bus abstraction.

So I answered your question? If so, it was by mistake, because I still
don't exactly understand your point (although I would like to).
The difference is that we don't register a new platform device for the
mii_bus (we use mii_bus->dev = ocelot->dev), and that it's not in
drivers/net/phy/mdio-fsl-enetc.c, although it could be, but I don't
see a clear benefit.

To go back to the original point, I suppose the answer is no: we did not
consider using a mii_bus abstraction and in hindsight, we do not see the
point but it was interesting you brought that up. Case closed from my
side, thank you.
-- 
Florian