[Linux-stm32] [PATCH RFC net-next 6/7] net: stmmac: add helpers to indicate WoL enable status

[Gatien CHEVALLIER]

On 7/28/25 19:54, Russell King (Oracle) wrote:
> On Mon, Jul 28, 2025 at 07:28:01PM +0200, Andrew Lunn wrote:
>>> +static inline bool stmmac_wol_enabled_mac(struct stmmac_priv *priv)
>>> +{
>>> +	return priv->plat->pmt && device_may_wakeup(priv->device);
>>> +}
>>> +
>>> +static inline bool stmmac_wol_enabled_phy(struct stmmac_priv *priv)
>>> +{
>>> +	return !priv->plat->pmt && device_may_wakeup(priv->device);
>>> +}
>>
>> I agree this is a direct translation into a helper.
>>
>> Reviewed-by: Andrew Lunn <andrew at lunn.ch>
>>
>> I'm guessing at some point you want to change these two
>> helpers. e.g. at some point, you want to try getting the PHY to do the
>> WoL, independent of !priv->plat->pmt?
>>
>>> -	if (device_may_wakeup(priv->device) && !priv->plat->pmt)
>>> +	if (stmmac_wol_enabled_phy(priv))
>>>   		phylink_speed_down(priv->phylink, false);
>>
>> This might be related to the next patch. But why only do speed down
>> when PHY is doing WoL? If the MAC is doing WoL, you could also do a
>> speed_down.
> 
> No idea, but that's what the code currently does, and, as ever with
> a cleanup series, I try to avoid functional changes in cleanup series.
> 
> Also, bear in mind that I can't test any of this.
> 
> We haven't yet been successful in getting WoL working in mainline. It
> _seems_ that the Jetson Xaiver NX platform should be using PHY mode,
> but the Realtek PHY driver is definitely broken for WoL. Even with
> that hacked, and along with other fixes that I've been given, I still
> can't get the SoC to wake up via WoL. In fact, the changes to change
> DT to specify the PHY interrupt as being routed through the PM
> controller results in normal PHY link up/down interrupts no longer
> working.
> 
> I'd like someone else to test functional changes!
> 

Hello Russel,

First of all, thank you for taking the time to improve this code.
What exactly do you mean by hacking? Forcing !priv->plat->pmt?