Re: [PATCH v2 06/10] net: pcs: xpcs: add ANRESTART support for SGMII link recovery

[Coia Prant <coiaprant@gmail.com>]

On August 2, 2026 10:30:52 PM GMT+08:00, Andrew Lunn <andrew@lunn.ch> wrote:
>On Sun, Aug 02, 2026 at 11:11:40AM +0800, Coia Prant wrote:
>> Hi Andrew,
>> 
>> Thanks for the careful review.
>> 
>> > This probably needs more wide spread testing. I would be afraid that
>> > if you kick off an AN cycle, you loose the link until it completes?
>> > You would of noticed that on your hardware, but other hardware
>> > obviously behaves differently...
>> 
>> The trigger condition is: link is down AND an_complete is set.
>
>Ah, i missed the "link is down" check. That makes it safer.
>
>I still find it interesting that this devices is different. Has the
>silicon vendor done the integration of the licensed IP wrong?
>
>> Regarding testing:
>> - Jiawen Wu has tested this on Wangxun NICs and reported no regression
>>   (Tested-by tag included in the patch).
>
>Probably Maxime will test on the platform he has, but i don't know if
>any use the XPCS.
>
>	Andrew

Hi Andrew,

Thanks for the follow-up.

Honestly, I'm not sure why the RK3568 XPCS behaves this way. I don't have
other DW XPCS hardware to compare against, so I can't say whether this is
a silicon integration quirk or something else.

What I can confirm is that on RK3568, the PCS does not restart AN
automatically when the link drops and comes back. Without this ANRESTART
change, the link stays dead forever.

I agree that wider testing is needed. If Maxime or anyone else finds a
regression on other hardware, we can revisit and condition it with a
platform-specific flag. For now, the change is limited to the SGMII
recovery path with a safe condition (link down + an_complete), so the
risk should be minimal.

Thanks,
Coia