On Mon, Jan 25, 2016 at 01:37:31PM +0100, Lothar Waßmann wrote:
> On Fri, 22 Jan 2016 16:26:10 +0000 Mark Brown wrote:

> > You're not just making the interrupt optional, you are also implementing
> > polling support.  That's really unusual and there's no clear reason for

> Any suggestions how to handle this case in a more sensible way?

The above, for example - make the interrupt optional.