Re: [PATCH net-next v1 3/4] net: phy: do not force EEE support

[Andrew Lunn]

On Mon, Feb 20, 2023 at 02:56:04PM +0100, Oleksij Rempel wrote:
> With following patches:
> commit 9b01c885be36 ("net: phy: c22: migrate to genphy_c45_write_eee_adv()")
> commit 5827b168125d ("net: phy: c45: migrate to genphy_c45_write_eee_adv()")
> 
> we set the advertisement to potentially supported values. This behavior
> may introduce new regressions on systems where EEE was disabled by
> default (BIOS or boot loader configuration or by other ways.)
> 
> At same time, with this patches, we would overwrite EEE advertisement
> configuration made over ethtool.
> 
> To avoid this issues, we need to cache initial and ethtool advertisement
> configuration and store it for later use.

This is good. I started adding phy_supports_eee() which MAC drivers
can call to say they support EEE. To help with that i also added
advertising_eee, which i initialise to 0. I've not get all the code in
place, but i was thinking to populate advertising_eee with
supported_eee when phy_supports_eee() is called, and add
genphy_c45_an_config_eee_aneg() or similar to be called during
phy_start(). So it look like you have implemented more than 1/2 of
what i wanted to do :-)

     Andrew