Re: [RFC] regmap: Add regmap_field APIs

[Srinivas KANDAGATLA <srinivas.kandagatla@st.com>]

On 10/06/13 10:15, Mark Brown wrote:
> On Sun, Jun 09, 2013 at 06:00:19PM +0200, Lars-Peter Clausen wrote:
> 
>>> +int regmap_field_write(struct regmap_field *field, unsigned int val)
>>> +{
>>> +	int field_bits;
>>> +	unsigned int reg_mask;
>>> +	field_bits = field->msb - field->lsb + 1;
>>> +	reg_mask = ((BIT(field_bits) - 1) << field->lsb);
>>> +	return regmap_update_bits(field->regmap, field->reg,
>>> +				reg_mask, val << field->lsb);
> 
>> Considering that you'd do the same calculations over and over again it would
>> probably make more sense store the mask rather than the msb in the struct
> 
> However as an interface for registering either is OK - the current
> MSB/LSB approach is probably better as that's what datasheets tend to
> include (which is why I didn't say anything).

regmap field interface still will be of lsb/msb style, However, As Lars
said, the internal data structure which holds these info can have mask
field rather than storing lsb/msb info.
I just posted a V2 patch with your review comments. I did not read your
response before I hit the send button.