Re: [PATCH net-next v3 1/2] ethtool: provide UAPI for PHY Signal Quality Index (SQI)

[Michal Kubecek]

On Wed, May 20, 2020 at 08:29:14AM +0200, Oleksij Rempel wrote:
> Signal Quality Index is a mandatory value required by "OPEN Alliance
> SIG" for the 100Base-T1 PHYs [1]. This indicator can be used for cable
> integrity diagnostic and investigating other noise sources and
> implement by at least two vendors: NXP[2] and TI[3].
> 
> [1] http://www.opensig.org/download/document/218/Advanced_PHY_features_for_automotive_Ethernet_V1.0.pdf
> [2] https://www.nxp.com/docs/en/data-sheet/TJA1100.pdf
> [3] https://www.ti.com/product/DP83TC811R-Q1
> 
> Signed-off-by: Oleksij Rempel <o.rempel@xxxxxxxxxxxxxx>
> ---

This looks good to me, there is just one thing I'm not sure about:

> diff --git a/include/linux/phy.h b/include/linux/phy.h
> index 59344db43fcb1..950ba479754bd 100644
> --- a/include/linux/phy.h
> +++ b/include/linux/phy.h
> @@ -706,6 +706,8 @@ struct phy_driver {
>  			    struct ethtool_tunable *tuna,
>  			    const void *data);
>  	int (*set_loopback)(struct phy_device *dev, bool enable);
> +	int (*get_sqi)(struct phy_device *dev);
> +	int (*get_sqi_max)(struct phy_device *dev);
>  };
>  #define to_phy_driver(d) container_of(to_mdio_common_driver(d),		\
>  				      struct phy_driver, mdiodrv)

I'm not sure if it's a good idea to define two separate callbacks. It
means adding two pointers instead of one (for every instance of the
structure, not only those implementing them), doing two calls, running
the same checks twice, locking twice, checking the result twice.

Also, passing a structure pointer would mean less code changed if we
decide to add more related state values later.

What do you think?

If you don't agree, I have no objections so

Reviewed-by: Michal Kubecek <mkubecek@xxxxxxx>

Michal

Attachment: signature.asc
Description: Digital signature