On Mon, Aug 06, 2018 at 03:39:53PM +0000, Popa, Stefan Serban wrote:
> On Lu, 2018-08-06 at 15:39 +0100, Mark Brown wrote:

> > I suggested having an explicit operation to check if a register supports
> > this mode of operation when I reviewed an earlier version, I didn't
> > notice a reply on that?

> Sorry, I got confused about what needs to be done. Should I add a new field
> to the regmap_config struct which will be checked during the function call?
> Something like can_multi_write? How do you suggest it should be called? Is
> readable_noinc_reg ok?

Like I said last time I'd suggest making it more like the readable and
volatile checks - I'm fairly sure I've seen devices that don't
autoincrement only for some registers.