<Query> Reg: Adding a gpio initialization sequence in dts is valid or not

Stephen Warren swarren at wwwdotorg.org
Thu Aug 9 01:38:09 EST 2012


On 08/08/2012 07:30 AM, Rob Herring wrote:
> On 08/08/2012 04:52 AM, Leela Krishna Amudala wrote:
>> Hello All,
>>
>> I came across that to set power to an LCD, few panels require a pull
>> up/pull down operation on single GPIO line
>> and few panels need some gpio line setting sequence with some delay in between.
...
>> Putting the sequence in dts file like below for an LCD which requires
>> gpio sequence
>>                lcd-reset-gpio = <&gpx1 5 1 3 0>,    (3 - for pull up)
>>                                       <&gpx1 5 1 1 0>,    (1 - for pull down)
>>                                       <&gpx1 5 1 3 0>;    (3 - for pull up)
>>                delay = x;
>>
>> and only one gpio entry in the case of LCD which doesn't require gpio
>> sequence setting
>>                lcd-reset-gpio = <&gpx1 5 1 3 0>,    (3 - pull up)
>>
>> The 4 arguments in the above handle denotes
>>
>> <[phandle of the gpio controller node]
>>       [pin number within the gpio controller]
>>       [mux function]
>>       [pull up/down]
>>       [drive strength]>

Mux function, pull-up/down and drive-strength should be handled by the
pinctrl subsystem, and associated DT bindings.

I suppose the proposal in the "runtime interpreted power sequences"
thread could be extended to have an action for activating a particular
pinctrl state too.


More information about the devicetree-discuss mailing list