Re: [PATCH v2 8/8] ARM: dts: qcom: Add Samsung Galaxy S4

["MINETTE Alexandre" <contact@alex-min.fr>]

Ah yes, I see. I tried GSBI for both the MUIC/extcon bus and the LED bus,
but it does not work on my GT-I9505: USB does not enumerate with the MUIC
on GSBI2, and the LED on GSBI3 times out when setting brightness (-110, with no visble changes).
So I think these should stay as i2c-gpio like downstream.
-- 
  MINETTE Alexandre
  contact@alex-min.fr

Lun 04 mai 2026, à 09:41, Konrad Dybcio a écrit :
> On 5/4/26 9:40 AM, Konrad Dybcio wrote:
>> On 5/1/26 9:41 AM, MINETTE Alexandre wrote:
>>> Thanks a lot for the suggestion, I checked the downstream board files
>>> and it does not look like these buses are wired for GSBI.
>>>
>>> There, the MAX77693 bus is instantiated as a dedicated i2c-gpio bus on
>>> GPIO 22/23, and the AN30259A LED bus as a dedicated i2c-gpio bus on
>>> GPIO 6/7. The native APQ8064 GSBI2/GSBI3 I2C pins are different
>>> (GPIO 24/25 and GPIO 8/9), so these two buses do not seem to be wired
>>> to the GSBI controllers on jflte.
>> 
>> What I meant is that the pins that the i2c-gpio nodes reference
>> are actually wired (inside the SoC) to the GSBI I2C controllers.
>> 
>> You'll notice that in drivers/pinctrl/qcom/pinctrl-apq8064.c,
>> there are bits like:
>> 
>> PINGROUP(6, gsbi3, NA, NA, NA, NA, NA, NA, NA, NA, NA),
>> 
>> which mean "function 0 is GPIO [implicitly in the macro], function
>> 1 is mux to the GSBI controller"
>
> Unless you meant "these are indeed GSBI pins, but within that GSBIn
> pingroup, those specific ones aren't assigned to SDA/SCL when
> proto==I2C, to which I'm afraid I won't be able to find docs for (too
> old)
>
> Konrad