[PATCH net-next v16 02/15] net: ethtool: Introduce ETHTOOL_LINK_MEDIUM_* values

[Maxime Chevallier]

Hi Jakub,

On 19/11/2025 04:15, Jakub Kicinski wrote:
> On Thu, 13 Nov 2025 09:14:04 +0100 Maxime Chevallier wrote:
>> diff --git a/include/linux/ethtool.h b/include/linux/ethtool.h
>> index c2d8b4ec62eb..ad2b5ed9522b 100644
>> --- a/include/linux/ethtool.h
>> +++ b/include/linux/ethtool.h
>> @@ -216,13 +216,26 @@ static inline u8 *ethtool_rxfh_context_key(struct ethtool_rxfh_context *ctx)
>>  void ethtool_rxfh_context_lost(struct net_device *dev, u32 context_id);
>>  
>>  struct link_mode_info {
>> -	int                             speed;
>> -	u8                              lanes;
>> -	u8                              duplex;
>> +	int	speed;
>> +	u8	lanes;
>> +	u8	min_pairs;
>> +	u8	pairs;
>> +	u8	duplex;
>> +	u16	mediums;
>>  };
>>  
>>  extern const struct link_mode_info link_mode_params[];
>>  
>> +extern const char ethtool_link_medium_names[][ETH_GSTRING_LEN];
>> +
>> +static inline const char *phy_mediums(enum ethtool_link_medium medium)
>> +{
>> +	if (medium >= __ETHTOOL_LINK_MEDIUM_LAST)
>> +		return "unknown";
>> +
>> +	return ethtool_link_medium_names[medium];
>> +}
> 
> Will this function be called by a lot of drivers? Would be great to
> find a more suitable place for it, ethtool.h is included by thousands
> of objects :(

I expect this to be mostly used by the current DT parsing code, as well
as a few print statements in the core. So indeed let's move this
somewhere more specific.

> 
>>  /* declare a link mode bitmap */
>>  #define __ETHTOOL_DECLARE_LINK_MODE_MASK(name)		\
>>  	DECLARE_BITMAP(name, __ETHTOOL_LINK_MODE_MASK_NBITS)
>> diff --git a/include/uapi/linux/ethtool.h b/include/uapi/linux/ethtool.h
>> index 8bd5ea5469d9..6ed235053aed 100644
>> --- a/include/uapi/linux/ethtool.h
>> +++ b/include/uapi/linux/ethtool.h
>> @@ -2587,4 +2587,24 @@ enum phy_upstream {
>>  	PHY_UPSTREAM_PHY,
>>  };
>>  
>> +enum ethtool_link_medium {
>> +	ETHTOOL_LINK_MEDIUM_BASET = 0,
>> +	ETHTOOL_LINK_MEDIUM_BASEK,
>> +	ETHTOOL_LINK_MEDIUM_BASES,
>> +	ETHTOOL_LINK_MEDIUM_BASEC,
>> +	ETHTOOL_LINK_MEDIUM_BASEL,
>> +	ETHTOOL_LINK_MEDIUM_BASED,
>> +	ETHTOOL_LINK_MEDIUM_BASEE,
>> +	ETHTOOL_LINK_MEDIUM_BASEF,
>> +	ETHTOOL_LINK_MEDIUM_BASEV,
>> +	ETHTOOL_LINK_MEDIUM_BASEMLD,
>> +	ETHTOOL_LINK_MEDIUM_NONE,
>> +
>> +	__ETHTOOL_LINK_MEDIUM_LAST,
>> +};
> 
> Why is this in uAPI? I'd have expected it to either exist in the YAML
> spec if there's a new uAPI that needs it, or in kernel headers.
> Let's move it out for now to be safe, we can always move it in.

I originally planned to report that in linkmodes et.al. but changed my
mind, let me move that out as well.

> 
>> +#define ETHTOOL_MEDIUM_FIBER_BITS (BIT(ETHTOOL_LINK_MEDIUM_BASES) | \
>> +				   BIT(ETHTOOL_LINK_MEDIUM_BASEL) | \
>> +				   BIT(ETHTOOL_LINK_MEDIUM_BASEF))
> 
> Ditto.

Thanks for taking a look,

Maxime