Re: [PATCHv13][ 2/4] video: imxfb: Also add pwmr for the device tree.
Alexander Shiyan
shc_work at mail.ru
Fri Dec 6 04:40:36 EST 2013
> > > On Fri, Dec 06, 2013 at 12:03:54PM +0400, Alexander Shiyan wrote:
> > ...
> > > > > > - fsl,dmacr: DMA Control Register value. This is optional. By default, the
> > > > > > register is not modified as recommended by the datasheet.
> > > > > > +- fsl,pwmr: LCDC PWM Contrast Control Register value. That property is
> > > > > > + optional, but defining it is necessary to get the backlight working. If that
> > > > > > + property is ommited, the register is zeroed.
> > > > >
> > > > > Why isn't this implemented as a backlight driver? Static devicetree
> > > > > provided values is very limiting.
> > > >
> > > > Let's understand the terminology.
> > > > This register should be renamed according to the datasheet, i.e. LPCCR.
> > > > As I pointed out earlier, it is NOT control the backlight, this is a contrast control.
> > > > Yes, it works as PWM, but nothing do with the backlight subsystem.
> > > > Yes, we can make a driver for this PWM, but how are we going to control it?
> > > > I misunderstood something?
> > >
> > > I stumbled upon 'get the backlight working' which implied for me that it
> > > should be a backlight driver. But you're right and now I remember we
> > > talked about this already.
> >
> > Hallelujah.
> >
> > > I still think this should be something adjustable, not static data.
> > > Maybe we could change the wording to something like "This property
> > > provides the default value for the contrast control register" since even
> > > if we add driver support for controlling the contrast we still want
> > > to have a sane default.
> >
> > Sounds good.
> >
> > > BTW the contrast could be controlled with a lcd_device (see
> > > lcd_device_register) which seems to be very easy to implement.
> >
> > Address of register is placed within LCD area, so we cannot use this
> > memory region, I think is no so easy as you say....
>
> We do not need a separate driver for this. Look for example at
> drivers/video/bf537-lq035.c, it just calls lcd_device_register()
> in its probe function.
Nice. Seems this example even can handle LCD power
regulator from "[1/4] video: imxfb: Introduce regulator support."
---
More information about the linux-arm-kernel
mailing list