Re: [PATCH 2/2] ARM: DTS: Add devicetree file for the Galaxy S2

[Marek Szyprowski]

Hi

On 2020-03-31 17:29, PaweÅ Chmiel wrote:
> On Tue, 2020-03-31 at 15:55 +0200, Marek Szyprowski wrote:
>> Hi Paul,
>>
>> On 2020-03-31 15:09, Paul Cercueil wrote:
>>> Cc: PaweÅ Chmiel <pawel.mikolaj.chmiel@xxxxxxxxx>
>>>
>>> Hi Marek,
>>>
>>> Le mar. 31 mars 2020 Ã 7:36, Marek Szyprowski
>>> <m.szyprowski@xxxxxxxxxxx> a Ãcrit :
>>>> Hi Paul,
>>>>
>>>> On 2020-03-18 15:25, Paul Cercueil wrote:
>>>>>>>    +    };
>>>>>>>    +
>>>>>>>    +    tsp_reg: regulator-1 {
>>>>>>>    +        compatible = "regulator-fixed";
>>>>>>>    +        regulator-name = "TSP_FIXED_VOLTAGES";
>>>>>>>    +        regulator-min-microvolt = <3300000>;
>>>>>>>    +        regulator-max-microvolt = <3300000>;
>>>>>>>    +        gpio = <&gpl0 3 GPIO_ACTIVE_HIGH>;
>>>>>>>    +        startup-delay-us = <70000>;
>>>>>>>    +        enable-active-high;
>>>>>>>    +        regulator-boot-on;
>>>>>>>    +        regulator-always-on;
>>>>>>   always-on and boot-on should not be needed. You have a consumer
>>>>>> for this
>>>>>>   regulator.
>>>>>   About this: the touchscreen driver does not use a regulator, so I
>>>>>   believe that's why these properties were here.
>>>>>
>>>>>   I sent patches upstream to address the issue:
>>>>>   https://protect2.fireeye.com/url?k=e8aedc29-b53072b3-e8af5766-0cc47a336fae-759579fd576d8382&u=https://lkml.org/lkml/2020/3/15/94
>>>>>
>>>>>
>>>>>   I believe this means I cannot merge the i9100 devicetree until it is
>>>>>   acked.
>>>> One more information - similar change has been already posted, but it
>>>> looks it got lost then: https://patchwork.kernel.org/patch/10550903/
>>> I was aware of this patch, but didn't know it was sent upstream.
>>>
>>> This other patch uses two regulators, vdd/avdd but doesn't give any
>>> reason why.
>>>
>> I've checked the UniversalC210 schematic, which uses the same
>> touchscreen chip. There are 2 supplies to the touchscreen chip: 2.8V VDD
>> and 3.3V AVDD. Both are enabled by the same GPIO pin though. There is
>> however no reset GPIO pin there.
> Hi
> Don't remember now how it worked on Galaxy S1, but it looks like it has
> the same setup - two regulators enabled by one GPIO pin.

It is quite common for the regulators to share the enable GPIO line, so 
this is not an issue. Regulator framework supports such configuration 
for ages. I'm curious about the reset GPIO line, which were made 
mandatory for the regulators control in your patch. I didn't find it on 
any schematic of the devices with this touch screen, but I don't have a 
schematic of the Galaxy S1. IMHO the reset signal is simply derived from 
the supply regulators with some simple circuit, but I don't have 
schematic of that part.

Best regards
-- 
Marek Szyprowski, PhD
Samsung R&D Institute Poland