[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <5d16e7c3201df22074019574d947dab1b5934b87@linux.dev>
Date: Fri, 23 May 2025 02:10:00 +0000
From: "Yajun Deng" <yajun.deng@...ux.dev>
To: "Andrew Lunn" <andrew@...n.ch>
Cc: hkallweit1@...il.com, linux@...linux.org.uk, davem@...emloft.net,
edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com,
netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net-next] net: phy: Synchronize c45_ids to phy_id
May 22, 2025 at 9:39 PM, "Andrew Lunn" <andrew@...n.ch> wrote:
>
> On Thu, May 22, 2025 at 09:19:18PM +0800, Yajun Deng wrote:
>
> >
> > The phy_id_show() function emit the phy_id for the phy device. If the phy
> >
> > device is a c45 device, the phy_id is empty. In other words, phy_id_show()
> >
> > only works with the c22 device.
> >
> >
> >
> > Synchronize c45_ids to phy_id, phy_id_show() will work with both the c22
> >
> > and c45 devices.
> >
>
> First off, they are different things. A device can have both a C22 and
>
> a collection of C45 IDs. So they should not be mixed up in one sysfs
>
> attribute.
In get_phy_device(), only one of get_phy_c45_ids() or get_phy_c22_id()
is called. Even if a device has both a c22 and c45 IDs, only one of the
phy_id and the c45_ids has values.
>
> The second point has already been made by Russell, there are lots of
>
> different C45 IDs, and phylib will try to find a driver based on any
>
> of them.
I noticed that. I tested the BCM89890, 88X3310 and 88Q2110 PHY devices,
and the ID is always the same in different MMDs.
>
> If you want to export the C45 IDs, please think about adding new sysfs
>
> attributes.
>
Okay.
> Andrew
>
Powered by blists - more mailing lists