[PATCH net-next v21 02/14] net: ethtool: Introduce ETHTOOL_LINK_MEDIUM_* values

[Paolo Abeni]

On 11/29/25 9:22 AM, Maxime Chevallier wrote:
> @@ -298,138 +321,149 @@ static_assert(ARRAY_SIZE(link_mode_names) == __ETHTOOL_LINK_MODE_MASK_NBITS);
>  		.speed	= SPEED_UNKNOWN, \
>  		.lanes	= 0, \
>  		.duplex	= DUPLEX_UNKNOWN, \
> +		.mediums = BIT(ETHTOOL_LINK_MEDIUM_NONE), \
>  	}
>  
>  const struct link_mode_info link_mode_params[] = {
> -	__DEFINE_LINK_MODE_PARAMS(10, T, Half),
> -	__DEFINE_LINK_MODE_PARAMS(10, T, Full),
> -	__DEFINE_LINK_MODE_PARAMS(100, T, Half),
> -	__DEFINE_LINK_MODE_PARAMS(100, T, Full),
> -	__DEFINE_LINK_MODE_PARAMS(1000, T, Half),
> -	__DEFINE_LINK_MODE_PARAMS(1000, T, Full),
> +	__DEFINE_LINK_MODE_PARAMS_PAIRS(10, T, 2, 4, Half, T),
> +	__DEFINE_LINK_MODE_PARAMS_PAIRS(10, T, 2, 4, Full, T),
> +	__DEFINE_LINK_MODE_PARAMS_PAIRS(100, T, 2, 4, Half, T),
> +	__DEFINE_LINK_MODE_PARAMS_PAIRS(100, T, 2, 4, Full, T),
> +	__DEFINE_LINK_MODE_PARAMS_PAIRS(1000, T, 4, 4, Half, T),
> +	__DEFINE_LINK_MODE_PARAMS_PAIRS(1000, T, 4, 4, Full, T),
>  	__DEFINE_SPECIAL_MODE_PARAMS(Autoneg),
>  	__DEFINE_SPECIAL_MODE_PARAMS(TP),
>  	__DEFINE_SPECIAL_MODE_PARAMS(AUI),
>  	__DEFINE_SPECIAL_MODE_PARAMS(MII),
>  	__DEFINE_SPECIAL_MODE_PARAMS(FIBRE),
>  	__DEFINE_SPECIAL_MODE_PARAMS(BNC),
> -	__DEFINE_LINK_MODE_PARAMS(10000, T, Full),
> +	__DEFINE_LINK_MODE_PARAMS_PAIRS(10000, T, 4, 4, Full, T),
>  	__DEFINE_SPECIAL_MODE_PARAMS(Pause),
>  	__DEFINE_SPECIAL_MODE_PARAMS(Asym_Pause),
> -	__DEFINE_LINK_MODE_PARAMS(2500, X, Full),
> +	__DEFINE_LINK_MODE_PARAMS_MEDIUMS(2500, X, Full,
> +					  __MED(C) | __MED(S) | __MED(L)),
>  	__DEFINE_SPECIAL_MODE_PARAMS(Backplane),
> -	__DEFINE_LINK_MODE_PARAMS(1000, KX, Full),
> -	__DEFINE_LINK_MODE_PARAMS(10000, KX4, Full),
> -	__DEFINE_LINK_MODE_PARAMS(10000, KR, Full),
> +	__DEFINE_LINK_MODE_PARAMS(1000, KX, Full, K),
> +	__DEFINE_LINK_MODE_PARAMS(10000, KX4, Full, K),
> +	__DEFINE_LINK_MODE_PARAMS(10000, KR, Full, K),
>  	[ETHTOOL_LINK_MODE_10000baseR_FEC_BIT] = {
>  		.speed	= SPEED_10000,
>  		.lanes	= 1,
>  		.duplex = DUPLEX_FULL,

The AI review points that medium is not initialized here:

https://netdev-ai.bots.linux.dev/ai-review.html?id=437cd013-c6a6-49e1-bec1-de4869930c7a#patch-1

Is that intentional? It should deserve at least an explanation in the
commit message.

Somewhat related, AI raised on the first patch the same question raised
on a previous iteration, and I assumed you considered that valid,
according to:

https://lore.kernel.org/netdev/f753719e-2370-401d-a001-821bdd5ee838@bootlin.com/

Otherwise I think some wording in the commit message explaining why the
AI feedback is incorrect would be useful.

/P