Re: [PATCH net-next] net: phy: dp83822: Add support for PHY LEDs on DP83822

[Dimitri Fedrau]

Am Tue, Dec 17, 2024 at 06:13:25PM +0100 schrieb Andrew Lunn:
> > +static int dp83822_led_hw_control_set(struct phy_device *phydev, u8 index,
> > +				      unsigned long rules)
> > +{
> > +	int mode;
> > +
> > +	mode = dp83822_led_mode(index, rules);
> > +	if (mode < 0)
> > +		return mode;
> > +
> > +	if (index == DP83822_LED_INDEX_LED_0 || index == DP83822_LED_INDEX_COL_GPIO2)
> > +		return phy_modify_mmd(phydev, MDIO_MMD_VEND2,
> > +				      MII_DP83822_MLEDCR, DP83822_MLEDCR_CFG,
> > +				      FIELD_PREP(DP83822_MLEDCR_CFG, mode));
> > +	else if (index == DP83822_LED_INDEX_LED_1_GPIO1)
> > +		return phy_modify_mmd(phydev, MDIO_MMD_VEND2,
> > +				      MII_DP83822_LEDCFG1,
> > +				      DP83822_LEDCFG1_LED1_CTRL,
> > +				      FIELD_PREP(DP83822_LEDCFG1_LED1_CTRL,
> > +						 mode));
> > +	else
> > +		return phy_modify_mmd(phydev, MDIO_MMD_VEND2,
> > +				      MII_DP83822_LEDCFG1,
> > +				      DP83822_LEDCFG1_LED3_CTRL,
> > +				      FIELD_PREP(DP83822_LEDCFG1_LED3_CTRL,
> > +						 mode));
> 
> index is taken direct from DT. Somebody might have:
> 
>            leds {
>                 #address-cells = <1>;
>                 #size-cells = <0>;
> 
>                 led@42 {
>                     reg = <42>;
>                     color = <LED_COLOR_ID_WHITE>;
>                     function = LED_FUNCTION_LAN;
>                     default-state = "keep";
>                 };
>             };
> 
> so you should not assume if it is not 0, 1 or 2, then it must be
> 3. Please always validate index.
>

By the way. Wouldn't it be helpful adding a u32 max_leds to
struct phy_driver ? Every driver supporting PHY LEDs validates index at the
moment. With max_leds it should be easy to check it in of_phy_leds and
return with an error if index is not valid.

Best regards,
Dimitri