Re: [PATCH 2/2] net:ethernet:enic: map ethtool link modes by VIC type

[Andrew Lunn]

> +	switch (sub_dev_id) {
> +	case PCI_SUBDEV_ID_CISCO_VIC_1225:
> +	case PCI_SUBDEV_ID_CISCO_VIC_1227:
> +		ethtool_link_ksettings_add_link_mode(ecmd, supported,
> +						     10000baseSR_Full);
> +		ethtool_link_ksettings_add_link_mode(ecmd, advertising,
> +						     10000baseSR_Full);

Do these ports support autoneg? Is it really advertising a single link
mode to the link partner for the negotiation process?

>  static int enic_get_ksettings(struct net_device *netdev,
>  			      struct ethtool_link_ksettings *ecmd)
>  {
>  	struct enic *enic = netdev_priv(netdev);
>  	struct ethtool_link_settings *base = &ecmd->base;
>  
> -	ethtool_link_ksettings_add_link_mode(ecmd, supported,
> -					     10000baseT_Full);
> -	ethtool_link_ksettings_add_link_mode(ecmd, supported, FIBRE);
> -	ethtool_link_ksettings_add_link_mode(ecmd, advertising,
> -					     10000baseT_Full);
> -	ethtool_link_ksettings_add_link_mode(ecmd, advertising, FIBRE);
> -	base->port = PORT_FIBRE;
> +	enic_get_supp_adv_media_type(netdev, ecmd);
>  
>  	if (netif_carrier_ok(netdev)) {
>  		base->speed = vnic_dev_port_speed(enic->vdev);

Ah, if you look at the rest of enic_get_ksettings() you see:

	base->autoneg = AUTONEG_DISABLE;

So this hardware does not support autoneg. So all these advertising
values are pointless.

       Andrew