| Message ID | ad4b5c29-abbc-450c-bada-5f671c287325@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net-next] ixgene-v2: prepare for phylib stop exporting phy_10_100_features_array | expand |
> + linkmode_clear_bit(ETHTOOL_LINK_MODE_10baseT_Full_BIT, > + phydev->supported); > + linkmode_clear_bit(ETHTOOL_LINK_MODE_10baseT_Half_BIT, > + phydev->supported); > + linkmode_clear_bit(ETHTOOL_LINK_MODE_100baseT_Full_BIT, > + phydev->supported); > + linkmode_clear_bit(ETHTOOL_LINK_MODE_100baseT_Half_BIT, > + phydev->supported); > + linkmode_clear_bit(ETHTOOL_LINK_MODE_1000baseT_Half_BIT, > + phydev->supported); > + > + phy_advertise_supported(phydev); phy_remove_link_mode() would be better. The MAC driver then does not need to know about the insides of the phydev structure, and it implicitly handles this last part. Andrew --- pw-bot: cr
On 11.02.2025 22:24, Andrew Lunn wrote: >> + linkmode_clear_bit(ETHTOOL_LINK_MODE_10baseT_Full_BIT, >> + phydev->supported); >> + linkmode_clear_bit(ETHTOOL_LINK_MODE_10baseT_Half_BIT, >> + phydev->supported); >> + linkmode_clear_bit(ETHTOOL_LINK_MODE_100baseT_Full_BIT, >> + phydev->supported); >> + linkmode_clear_bit(ETHTOOL_LINK_MODE_100baseT_Half_BIT, >> + phydev->supported); >> + linkmode_clear_bit(ETHTOOL_LINK_MODE_1000baseT_Half_BIT, >> + phydev->supported); >> + >> + phy_advertise_supported(phydev); > > phy_remove_link_mode() would be better. The MAC driver then does not > need to know about the insides of the phydev structure, and it > implicitly handles this last part. > Only small drawback is that it unnecessarily calls phy_advertise_supported() after each statement. Therefore I decided against it. But you're right, it's cleaner. > Andrew > Heiner > --- > pw-bot: cr
diff --git a/drivers/net/ethernet/apm/xgene-v2/mdio.c b/drivers/net/ethernet/apm/xgene-v2/mdio.c index eba06831a..b41f15501 100644 --- a/drivers/net/ethernet/apm/xgene-v2/mdio.c +++ b/drivers/net/ethernet/apm/xgene-v2/mdio.c @@ -97,7 +97,6 @@ void xge_mdio_remove(struct net_device *ndev) int xge_mdio_config(struct net_device *ndev) { - __ETHTOOL_DECLARE_LINK_MODE_MASK(mask) = { 0, }; struct xge_pdata *pdata = netdev_priv(ndev); struct device *dev = &pdata->pdev->dev; struct mii_bus *mdio_bus; @@ -137,17 +136,18 @@ int xge_mdio_config(struct net_device *ndev) goto err; } - linkmode_set_bit_array(phy_10_100_features_array, - ARRAY_SIZE(phy_10_100_features_array), - mask); - linkmode_set_bit(ETHTOOL_LINK_MODE_1000baseT_Half_BIT, mask); - linkmode_set_bit(ETHTOOL_LINK_MODE_AUI_BIT, mask); - linkmode_set_bit(ETHTOOL_LINK_MODE_MII_BIT, mask); - linkmode_set_bit(ETHTOOL_LINK_MODE_FIBRE_BIT, mask); - linkmode_set_bit(ETHTOOL_LINK_MODE_BNC_BIT, mask); - - linkmode_andnot(phydev->supported, phydev->supported, mask); - linkmode_copy(phydev->advertising, phydev->supported); + linkmode_clear_bit(ETHTOOL_LINK_MODE_10baseT_Full_BIT, + phydev->supported); + linkmode_clear_bit(ETHTOOL_LINK_MODE_10baseT_Half_BIT, + phydev->supported); + linkmode_clear_bit(ETHTOOL_LINK_MODE_100baseT_Full_BIT, + phydev->supported); + linkmode_clear_bit(ETHTOOL_LINK_MODE_100baseT_Half_BIT, + phydev->supported); + linkmode_clear_bit(ETHTOOL_LINK_MODE_1000baseT_Half_BIT, + phydev->supported); + + phy_advertise_supported(phydev); pdata->phy_speed = SPEED_UNKNOWN; return 0;
As part of phylib cleanup we plan to stop exporting the feature arrays. So explicitly remove the modes not supported by the MAC. The media type bits don't have any impact on kernel behavior, so don't touch them. Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> --- drivers/net/ethernet/apm/xgene-v2/mdio.c | 24 +++++++++++----------- 3 files changed, 13 insertions(+), 14 deletions(-)