[PATCH net-next] net: dsa: mt7530: Add support for EEE features

[René van Dorst]

Quoting DENG Qingfang <dqfext at gmail.com>:

Hi Qingfang,

Thanks for the review.

> Hi René,
>
> On Sat, Apr 10, 2021 at 6:54 AM René van Dorst <opensource at vdorst.com> wrote:
>> --- a/drivers/net/dsa/mt7530.c
>> +++ b/drivers/net/dsa/mt7530.c
>> @@ -2568,6 +2568,11 @@ static void  
>> mt753x_phylink_mac_link_up(struct dsa_switch *ds, int port,
>>                         mcr |= PMCR_TX_FC_EN;
>>                 if (rx_pause)
>>                         mcr |= PMCR_RX_FC_EN;
>> +
>> +               if (mode == MLO_AN_PHY && phydev &&
>> +                   !(priv->eee_disabled & BIT(port)) &&
>> +                   phy_init_eee(phydev, 0) >= 0)
>> +                       mcr |= PMCR_FORCE_EEE1G | PMCR_FORCE_EEE100;
>
> You should adjust this according to e->advertised.

I now better understand EEE part in phylink.
I refactor the code a lot and also tested with different
port with various eee-broken-1000t and eee-broken-100tx set
in the device tree.

Looking at mcr value, in all the cases the bit are set right.

>
>>         }
>>
>>         mt7530_set(priv, MT7530_PMCR_P(port), mcr);
>> @@ -2800,6 +2805,49 @@ mt753x_phy_write(struct dsa_switch *ds, int  
>> port, int regnum, u16 val)
>>         return priv->info->phy_write(ds, port, regnum, val);
>>  }
>>
>> +static int mt753x_get_mac_eee(struct dsa_switch *ds, int port,
>> +                             struct ethtool_eee *e)
>> +{
>> +       struct mt7530_priv *priv = ds->priv;
>> +       u32 eeecr, pmsr;
>> +
>> +       e->eee_enabled = !(priv->eee_disabled & BIT(port));
>> +
>> +       if (e->eee_enabled) {
>> +               eeecr = mt7530_read(priv, MT7530_PMEEECR_P(port));
>> +               e->tx_lpi_enabled = !(eeecr & LPI_MODE_EN);
>> +               e->tx_lpi_timer = GET_LPI_THRESH(eeecr);
>> +               pmsr = mt7530_read(priv, MT7530_PMSR_P(port));
>> +               e->eee_active = e->eee_enabled && !!(pmsr & PMSR_EEE1G);
>
> eee_enabled and eee_active will be set in phy_ethtool_get_eee, no need
> to set them here.

Thanks for pointing that out.
I refactor the code so it only set/report the LPI settings.

I already sended v2 with these changes.

>
>> +       }
>> +
>> +       return 0;
>> +}
>> +
>> +static int mt753x_set_mac_eee(struct dsa_switch *ds, int port,
>> +                             struct ethtool_eee *e)
>> +{
>> +       struct mt7530_priv *priv = ds->priv;
>> +       u32 eeecr;
>> +
>> +       if (e->eee_enabled) {
>> +               if (e->tx_lpi_timer > 0xFFF)
>> +                       return -EINVAL;
>> +               priv->eee_disabled &= ~BIT(port);
>> +               eeecr = mt7530_read(priv, MT7530_PMEEECR_P(port));
>> +               eeecr &= ~(LPI_THRESH_MASK | LPI_MODE_EN);
>> +               if (!e->tx_lpi_enabled)
>> +                       /* Force LPI Mode without a delay */
>> +                       eeecr |= LPI_MODE_EN;
>> +               eeecr |= SET_LPI_THRESH(e->tx_lpi_timer);
>> +               mt7530_write(priv, MT7530_PMEEECR_P(port), eeecr);
>> +       } else {
>> +               priv->eee_disabled |= BIT(port);
>> +       }
>> +
>> +       return 0;
>> +}


Greats,

René