Re: [5/6] mvneta: implement SGMII-based in-band link state signaling

[Stas Sergeev <stsp@list.ru>]

09.07.2015 12:19, Thomas Petazzoni пишет:
> Sebastien, Stas,
> 
> On Thu, 9 Jul 2015 11:03:26 +0200 (CEST), Sebastien Rannou wrote:
>> On Wed, 8 Jul 2015, Stas Sergeev wrote:
>>
>>> What is there? A phy chip, or something else?
>>
>> It's "something else", there's a phy which aggregates 4xSGMIIs to
>> 1xQSGMII, we are on the media side here, the MAC side is connected
>> to the switch through QSGMII.
>>
>>> Perhaps some DT property should be added to explicitly
>>> enable the use of the inband status...
>>
>> Yes, that would be fine.
> 
> Isn't it a bit weird to need a new DT property for this? Shouldn't
> fixed-link always imply this inband status thing?
That's how it is currently implemented. I thought its safe.
But what if the device on the other end does not generate the
inband status? I think this device is doing the wrong thing,
but nevertheless we have a regression at hands.

Currently the link status cannot be specified for fixed-link,
at all.
What I am going to code up, is the new property, like this:

fixed-link {
  link = "up" | "down" | "auto";
};

"auto" will mean the inband status.
Looks like a simple solution.