[PATCH v3 02/16] mfd: da8xx-cfgchip: New header file for CFGCHIP registers.

Sekhar Nori nsekhar at ti.com
Mon Apr 11 01:23:57 PDT 2016

On Monday 28 March 2016 10:12 PM, Sergei Shtylyov wrote:
> Hello.
> On 03/28/2016 06:02 PM, David Lechner wrote:
>>>> +/* register offsets */
>>>> +#define CFGCHIP_REG(n)                (n * 4)
>>>> +#define CFGCHIP0_REG                CFGCHIP_REG(0)
>>>> +#define CFGCHIP1_REG                CFGCHIP_REG(1)
>>>> +#define CFGCHIP2_REG                CFGCHIP_REG(2)
>>>> +#define CFGCHIP3_REG                CFGCHIP_REG(3)
>>>> +#define CFGCHIP4_REG                CFGCHIP_REG(4)
>>>     Why not just use CFGCHIP_REG(n) directly?
>> I considered that, but I went this way because A) the TRM uses, for
>> example,
>> "CFGCHIP2", so I wanted to keep "CFGCHIP" and "2" together

IMO, this is not that big of an issue. Anyone reading should be able to
make out that CFGCHIP_REG(0) is same as CFGCHIP0 referred to in the TRM.

>    I'd just drop the _REG suffix.
>> and B) this tells
>> you how many CFGCHIP registers there are, i.e. there is no CFGCHIP5_REG.
>    You can tell that in a comment. Having a parametrized macro and using
> it to just #define more macros doesn't appeal to me at all...

Agree with Sergei, I don't prefer the additional #defines as well.


More information about the linux-arm-kernel mailing list