[PATCH] ARM: dts: imx: add Gateworks Ventana GW5904 support

Shawn Guo shawnguo at kernel.org
Tue Mar 14 23:31:41 PDT 2017


On Tue, Mar 14, 2017 at 08:05:03AM -0700, Tim Harvey wrote:
> >> +&ldb {
> >> +     status = "okay";
> >> +
> >> +     lvds-channel at 0 {
> >> +             fsl,data-mapping = "spwg";
> >> +             fsl,data-width = <18>;
> >> +             status = "okay";
> >> +
> >> +             display-timings {
> >> +                     native-mode = <&timing0>;
> >> +                     timing0: hsd100pxn1 {
> >> +                             clock-frequency = <65000000>;
> >> +                             hactive = <1024>;
> >> +                             vactive = <768>;
> >> +                             hback-porch = <220>;
> >> +                             hfront-porch = <40>;
> >> +                             vback-porch = <21>;
> >> +                             vfront-porch = <7>;
> >> +                             hsync-len = <60>;
> >> +                             vsync-len = <10>;
> >> +                     };
> >> +             };
> >> +     };
> >
> > Take a look at commit 4dc633e9b019 ("ARM: dts: sabrelite: use
> > simple-panel instead of display-timings for LVDS0"), and consider to use
> > simple-panel?
> 
> I haven't moved to simple-panel yet because I have bootloader code
> that allows choosing/altering display timings with the goal being
> users don't need to recompile their device-tree or kernel to use a
> display with different timings. It seems to me that moving to
> simple-panel would make this even more difficult as while the
> bootloader could find and alter the panel's compatible property (in
> the case the kernel has a supported simple-panel compiled in) it no
> longer has access to the raw timings (in case the kernel doesn't have
> a simple-panel driver built-in already).
> 
> I do like the way simple-panel combines display timings with
> backlight, power supplies, dc bus, and a gpio enable but it doesn't
> encapsulate touch controller or expose timings to device-tree for easy
> manipulation.
> 
> What are you thoughts on this?

I'm fine with it, as you have a reason for that.

> 
> >
> <snip>
> >> +
> >> +&pwm2 {
> >> +     pinctrl-names = "default";
> >> +     pinctrl-0 = <&pinctrl_pwm2>; /* MX6_DIO1 */
> >> +     status = "disabled";
> >> +};
> >> +
> >> +&pwm3 {
> >> +     pinctrl-names = "default";
> >> +     pinctrl-0 = <&pinctrl_pwm3>; /* MX6_DIO2 */
> >> +     status = "disabled";
> >> +};
> >
> > Why do you have these two devices but disable them?
> 
> This is because I have a bootloader configuration that allows the user
> to choose between GPIO and PWM for the non-backlight PWM pins. I
> should probably add a comment to those nodes specifying that firmware
> modifies the status property.

Ditto

Shawn



More information about the linux-arm-kernel mailing list