[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4894f014-88ed-227a-7563-e3bf3b16e00c@gmail.com>
Date: Mon, 11 May 2020 12:06:31 -0700
From: Florian Fainelli <f.fainelli@...il.com>
To: Russell King - ARM Linux admin <linux@...linux.org.uk>,
Julien Beraud <julien.beraud@...lia.com>
Cc: "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
Andrew Lunn <andrew@...n.ch>,
Heiner Kallweit <hkallweit1@...il.com>
Subject: Re: net: phylink: supported modes set to 0 with genphy sfp module
On 5/11/2020 11:29 AM, Russell King - ARM Linux admin wrote:
> On Mon, May 11, 2020 at 05:45:02PM +0200, Julien Beraud wrote:
>> Following commit:
>>
>> commit 52c956003a9d5bcae1f445f9dfd42b624adb6e87
>> Author: Russell King <rmk+kernel@...linux.org.uk>
>> Date: Wed Dec 11 10:56:45 2019 +0000
>>
>> net: phylink: delay MAC configuration for copper SFP modules
>>
>>
>> In function phylink_sfp_connect_phy, phylink_sfp_config is called before
>> phylink_attach_phy.
>>
>> In the case of a genphy, the "supported" field of the phy_device is filled
>> by:
>> phylink_attach_phy->phy_attach_direct->phy_probe->genphy_read_abilities.
>>
>> It means that:
>>
>> ret = phylink_sfp_config(pl, mode, phy->supported, phy->advertising);
>> will have phy->supported with no bits set, and then the first call to
>> phylink_validate in phylink_sfp_config will return an error:
>>
>> return phylink_is_empty_linkmode(supported) ? -EINVAL : 0;
>>
>> this results in putting the sfp driver in "failed" state.
>
> Which PHY is this?
Using the generic PHY with a copper SFP module does not sound like a
great idea because without a specialized PHY driver (that is, not the
Generic PHY driver) there is not usually much that can happen.
--
Florian
Powered by blists - more mailing lists