On Wed, Oct 14, 2020 at 10:23 PM Leizhen (ThunderTown) <thunder.leizhen@xxxxxxxxxx> wrote: > > > > On 2020/10/14 21:50, Rob Herring wrote: > > On Wed, Oct 14, 2020 at 09:29:26AM +0800, Leizhen (ThunderTown) wrote: > >> > >> > >> On 2020/10/14 1:32, Dan Murphy wrote: > >>> Zhen > >>> > >>> On 10/13/20 11:08 AM, Zhen Lei wrote: > >>>> The property name used in arch/arm/boot/dts/mmp2-olpc-xo-1-75.dts is > >>>> cmd-gpio. > >>>> > >>>> arch/arm/boot/dts/mmp2-olpc-xo-1-75.dts:235: > >>>> cmd-gpio = <&gpio 155 GPIO_ACTIVE_HIGH>; > >>>> > >>>> Signed-off-by: Zhen Lei <thunder.leizhen@xxxxxxxxxx> > >>>> --- > >>>> Documentation/devicetree/bindings/misc/olpc,xo1.75-ec.yaml | 6 +++--- > >>>> 1 file changed, 3 insertions(+), 3 deletions(-) > >>>> > >>>> diff --git a/Documentation/devicetree/bindings/misc/olpc,xo1.75-ec.yaml b/Documentation/devicetree/bindings/misc/olpc,xo1.75-ec.yaml > >>>> index b3c45c046ba5e37..c7a06a9650db2ed 100644 > >>>> --- a/Documentation/devicetree/bindings/misc/olpc,xo1.75-ec.yaml > >>>> +++ b/Documentation/devicetree/bindings/misc/olpc,xo1.75-ec.yaml > >>>> @@ -24,7 +24,7 @@ properties: > >>>> compatible: > >>>> const: olpc,xo1.75-ec > >>>> - cmd-gpios: > >>>> + cmd-gpio: > >>> > >>> Preference is gpios not gpio. But Rob H accept or reject > >> > >> Look at the search result below. It seems that the driver have not been merged into mainline. > > > > Yes, in drivers/platform/olpc/olpc-xo175-ec.c. > > > > Your mistake is the gpiod api takes just 'cmd' as the GPIO core handles > > both forms. > > OK, thanks for your information. I have found that it defined by gpio_suffixes[]. > > > > >> But the property name is really used as cmd-gpio at mmp2-olpc-xo-1-75.dts:235, I don't think > >> the mmp2-olpc-xo-1-75.dts can make a mistake. Otherwise, the driver will not work properly. > >> Meanwhile, Both names cmd-gpios and cmd-gpio seem to be in use. But I prefer cmd-gpio, after > >> all, only one gpio is assigned now. The motorola,cmd-gpios add "s" because it contains 3 gpio. > > > > The preference is it is always '-gpios' just like it's always > > 'interrupts' or 'clocks'. > > > > However, whether to change this is really up to the OLPC folks. Given > > the driver has always supported both forms, it should be okay to change > > the dts. Though there could be other users besides the kernel. > > If both "cmd-gpios" and "cmd-gpio" are supported, should we use enum to list both > of them in yaml? or use patternProperties? No, we pick one or the other. Given Lubomir is okay with a dts change, we should use just 'cmd-gpios'. Rob