On Tue, Nov 13, 2012 at 11:18:34PM +0530, Laxman Dewangan wrote:
> On Tuesday 13 November 2012 01:16 PM, Mark Brown wrote:

> >I'm not sure your wake_base setting is sane here - this is expected to
> >be a register.  If it is a register why is it not a #define?

> There is no wake register. Here I kept for the incerease the wake
> conut so that irq_set_irq_wake() can be called.

So you want to write to random registers in the chip instead...  that
doesn't seem like a terribly clever idea.

> I think it is also not correct after change
> regmap: irq: initialize all irqs to wake disabled

> I need to set the irq_wake(0 without the register.

I can't really parse any of the above, sorry.

> There is 2 sets of registers which need to be configure, MASK and
> LINE enable.

So what does the second one do?