Re: [PATCH 1/3] net: thunderx: Cleanup PHY probing code.

[David Daney]

On 03/11/2016 01:35 PM, Andrew Lunn wrote:
[...]
> How usable is the hardware without a PHY driver?

The hardware has always in the past, still does, and probably always 
will work fine without a PHY driver.  Link up/down are correctly handled.

> Is a better solution
> that your write a very minimal PHY driver?

No.  Nothing would be gained.

All we are trying to do, is allow for loading of 1G PHY drivers via the 
-EPROBE_DEFER mechanism while continuing to allow the 10G and 40G ports 
to function without a PHY driver.

Specifically, we are *not* attempting to solve the problem of 
re-architecting the kernel phy_device infrastructure so that it would be 
possible to write a Cortina PHYs driver.  Nor are we proposing that a 
Cortina PHY driver be written that would fit into the current 
infrastructure.

To this end, I still think the current patch takes the best approach.

Thanks,
David Daney