RE: [PATCH, net] phy: aquantia: Fix AN when higher speeds than 1G are not advertised

From: Claudiu Manoil <claudiu.manoil@nxp.com>
To: Andrew Lunn <andrew@lunn.ch>
Cc: Heiner Kallweit <hkallweit1@gmail.com>,
	Florian Fainelli <f.fainelli@gmail.com>,
	Russell King <linux@armlinux.org.uk>,
	"David S. Miller" <davem@davemloft.net>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
	Ondrej Spacek <ondrej.spacek@nxp.com>
Subject: RE: [PATCH, net] phy: aquantia: Fix AN when higher speeds than 1G are not advertised
Date: Sun, 12 Jun 2022 18:47:45 +0000	[thread overview]
Message-ID: <AM9PR04MB8397DE1463E47F5AF8574A5F96A89@AM9PR04MB8397.eurprd04.prod.outlook.com> (raw)
In-Reply-To: <YqUjUQLOHUo55egO@lunn.ch>

> -----Original Message-----
> From: Andrew Lunn <andrew@lunn.ch>
> Sent: Sunday, June 12, 2022 2:21 AM
> To: Claudiu Manoil <claudiu.manoil@nxp.com>
[...]
> Subject: Re: [PATCH, net] phy: aquantia: Fix AN when higher speeds than 1G are
> not advertised
> 
> On Sat, Jun 11, 2022 at 06:13:12PM +0000, Claudiu Manoil wrote:
> > > -----Original Message-----
> > > From: Andrew Lunn <andrew@lunn.ch>
> > > Sent: Saturday, June 11, 2022 6:00 PM
> > [...]
> > > Subject: Re: [PATCH, net] phy: aquantia: Fix AN when higher speeds than 1G are
> > > not advertised
> > >
> > > On Fri, Jun 10, 2022 at 11:40:37AM +0300, Claudiu Manoil wrote:
> > > > Even when the eth port is resticted to work with speeds not higher than 1G,
> > > > and so the eth driver is requesting the phy (via phylink) to advertise up
> > > > to 1000BASET support, the aquantia phy device is still advertising for 2.5G
> > > > and 5G speeds.
> > > > Clear these advertising defaults when requested.
> > >
> > > Hi Claudiu
> > >
> > > genphy_c45_an_config_aneg(phydev) is called in aqr_config_aneg, which
> > > should set/clear MDIO_AN_10GBT_CTRL_ADV5G and
> > > MDIO_AN_10GBT_CTRL_ADV2_5G. Does the aQuantia PHY not have these bits?
> > >
> >
> > Hi Andrew,
> >
> > Apparently it's not enough to clear the 7.20h register (aka MDIO_AN_10GBT_CTRL)
> > to stop AQR advertising for higher speeds, otherwise I wouldn't have bothered with
> > the patch.
> 
> I'm just trying to ensure we are not papering over a crack. I wanted
> to eliminate the possibility of a bug in that code which is changing
> 7.20h. If the datasheet for the aquantia states those bits are not
> used, then this patch is fine. If on the other hand the datasheet says
> 7.20 can be used to change the advertisement, we should investigate
> further what is really going on.
> 

I understand the situation is not ideal. What I could do is to provide some
logs showing that writing the correct configuration to 7.20h does not yield
the desired result, to have some sort of detailed evidence about the issue.
But I cannot do that until Tuesday at the earliest.
As for documentation, there's an App Note about configuring advertising
via the vendor specific 0xC400 reg but I don't think it's public. Not sure if
we can use that.
Meanwhile, it would be great if you or someone else could confirm the
issue on a different platform.

Thanks.