On Fri, Dec 27, 2024 at 12:13:40AM +0100, Tobias Jakobi wrote: > On 12/26/24 22:05, Guenter Roeck wrote: > > On Thu, Dec 26, 2024 at 06:00:19PM +0100, tjakobi@xxxxxxxxxxxxxxxxxxxxx wrote: > > > From: Tobias Jakobi <tjakobi@xxxxxxxxxxxxxxxxxxxxx> > > > > > > The driver is in full control of the enable mode, so we > > > don't need to read it from HW every single time. > > > > > > > That is not a reason for adding that much additional code. > > What is the problem that is being solved, and why is it worth that much > > additional code ? > I don't think it's that much additional code, but anyway: Reading from EC is > not exactly fast, and I want this value cached for another reason. It turns > out that some devices use a different scaling for the PWM value depending on > whether the PWM is controlled automatically by the EC, or manually through > the driver. And I don't want to do an additional EC read to figure this out, > if I can avoid it. Maybe it isn't that much code after the runtime feature checks are removed. Either case, since there are now two operations (reading/writing from/to the EC and caching the result), all associated operations will need to be mutex protected. Guenter
