[PATCH net-next v3 14/47] net: phy: aquantia: Add support for rate adaptation

Sean Anderson sean.anderson at seco.com
Sat Jul 16 15:45:01 PDT 2022


On 7/16/22 2:38 PM, Andrew Lunn wrote:
>> +#define VEND1_GLOBAL_CFG_10M			0x0310
>> +#define VEND1_GLOBAL_CFG_100M			0x031b
>> +#define VEND1_GLOBAL_CFG_1G			0x031c
>> +#define VEND1_GLOBAL_CFG_2_5G			0x031d
>> +#define VEND1_GLOBAL_CFG_5G			0x031e
>> +#define VEND1_GLOBAL_CFG_10G			0x031f
> 
> I completely read this wrong the first time... The common meaning of
> #defines line this is
> 
> VEND1_GLOBAL_CFG_ is the register and what follows indicates some bits
> in the register.
> 
> However, this is not true here, these are all registers. Maybe add
> _REG to the end? It makes them different to other defines for
> registers, but if i parsed it wrong, probably other will as well?

How about a comment like

/* The following registers all have similar layouts; first the registers... */
#define VEND1_GLOBAL_CFG_10M				0x0310
...
/* ... and now the fields */
#define VEND1_GLOBAL_CFG_RATE_ADAPT			GENMASK(8, 7)

>>   static int aqr107_read_rate(struct phy_device *phydev)
>>   {
>>   	int val;
>> +	u32 config_reg;
> 
> Revere Christmass tree. config_reg should be first.

OK

--Sean




More information about the linux-arm-kernel mailing list