RE: [BUG] net: phy: genphy_loopback: add link speed configuration

[Ismail, Mohammad Athari]

> -----Original Message-----
> From: Andrew Lunn <andrew@xxxxxxx>
> Sent: Wednesday, December 15, 2021 5:55 PM
> To: Ismail, Mohammad Athari <mohammad.athari.ismail@xxxxxxxxx>
> Cc: Oleksij Rempel <o.rempel@xxxxxxxxxxxxxx>; netdev@xxxxxxxxxxxxxxx;
> linux-kernel@xxxxxxxxxxxxxxx; Voon, Weifeng <weifeng.voon@xxxxxxxxx>;
> Wong, Vee Khee <vee.khee.wong@xxxxxxxxx>
> Subject: Re: [BUG] net: phy: genphy_loopback: add link speed configuration
> 
> > > -----Original Message-----
> > > From: Andrew Lunn <andrew@xxxxxxx>
> > > Sent: Wednesday, December 15, 2021 5:23 PM
> > > To: Ismail, Mohammad Athari <mohammad.athari.ismail@xxxxxxxxx>
> > > Cc: Oleksij Rempel <o.rempel@xxxxxxxxxxxxxx>;
> > > netdev@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; Voon, Weifeng
> > > <weifeng.voon@xxxxxxxxx>; Wong, Vee Khee
> <vee.khee.wong@xxxxxxxxx>
> > > Subject: Re: [BUG] net: phy: genphy_loopback: add link speed
> > > configuration
> > >
> > > > Thanks for the suggestion. The proposed solution also doesn't
> > > > work. Still
> > > get -110 error.
> > >
> > > Please can you trace where this -110 comes from. Am i looking at the
> > > wrong poll call?
> >
> > I did read the ret value from genphy_soft_reset() and
> phy_read_poll_timeout().
> > The -110 came from phy_read_poll_timeout().
> 
> O.K.
> 
> Does the PHY actually do loopback, despite the -110?

As Intel Elkhart Lake is using stmmac driver, in stmmac_selftest, return value of phy_loopback() is checked as well. If it return -110, the selftest that using PHY loopback will be recorded as -110 (fail).

> 
> I'm wondering if we should ignore the return value from
> phy_read_poll_timeout().

Removing/ignoring the return value from phy_read_poll_timeout() can work. But, the -110 error message will be displayed in dmesg. It is because there is phydev_err() as part of phy_read_poll_timeout() definition.

-Athari-

> 
> 	Andrew