On Mon, Oct 21, 2019 at 03:27:20PM +0300, Mika Westerberg wrote: > On Mon, Oct 14, 2019 at 11:43:48AM +0300, Andy Shevchenko wrote: > > Refactor restoring GPI_IE registers by using an introduced helper. > > + struct device *dev = pctrl->dev; > > const? It's only about dozen occurrences in 8 drivers altogether in the kernel to have const with struct device. So, I consider this pattern is unusual. > Also should we do the same here as we do with others and check first > whether we need to update the mask at all? I will look at it. I think we may do it as a separate change (this doesn't change any functionality). -- With Best Regards, Andy Shevchenko