Hello,
On Wed, 25 Sep 2024 16:01:28 -0700
Abhishek Chauhan <quic_abchauha@xxxxxxxxxxx> wrote:
> AQR115c reports incorrect PMA capabilities which includes
> 10G/5G and also incorrectly disables capabilities like autoneg
> and 10Mbps support.
>
> AQR115c as per the Marvell databook supports speeds up to 2.5Gbps
> with autonegotiation.
>
> Fixes: 0ebc581f8a4b ("net: phy: aquantia: add support for aqr115c")
> Link: https://lore.kernel.org/all/20240913011635.1286027-1-quic_abchauha@xxxxxxxxxxx/T/
> Signed-off-by: Abhishek Chauhan <quic_abchauha@xxxxxxxxxxx>
> ---
[...]
>
> +static int aqr115c_get_features(struct phy_device *phydev)
> +{
> + int ret;
> + __ETHTOOL_DECLARE_LINK_MODE_MASK(supported) = { 0, };
> +
> + /* Normal feature discovery */
> + ret = genphy_c45_pma_read_abilities(phydev);
> + if (ret)
> + return ret;
> +
> + /* PHY FIXUP */
> + /* Although the PHY sets bit 12.18.19.48, it does not support 5G/10G modes */
> + linkmode_clear_bit(ETHTOOL_LINK_MODE_10000baseT_Full_BIT, phydev->supported);
> + linkmode_clear_bit(ETHTOOL_LINK_MODE_10000baseKX4_Full_BIT, phydev->supported);
> + linkmode_clear_bit(ETHTOOL_LINK_MODE_10000baseKR_Full_BIT, phydev->supported);
> + linkmode_clear_bit(ETHTOOL_LINK_MODE_5000baseT_Full_BIT, phydev->supported);
> +
> + /* Phy supports Speeds up to 2.5G with Autoneg though the phy PMA says otherwise */
> + linkmode_copy(supported, phy_gbit_features);
> + linkmode_set_bit(ETHTOOL_LINK_MODE_2500baseX_Full_BIT, supported);
I still think you shouldn't report 2500BaseX, as you mentionned
it's a BaseT PHY. This is independent of the modes that the PHY uses to
connect to the MAC. Although the PHY can talk to the MAC using
2500BaseX on its MII interface, it looks like it can't use
2500BaseX on its MDI (the LP side of the PHY). There's the same issue in
patch 2.
Thanks,
Maxime
