Re: [PATCH v2 09/15] pinctrl: qcom: Add Kuno pinctrl driver

[Hardeep Sharma]

On 8/17/2026 8:55 PM, Konrad Dybcio wrote:
> On 8/17/26 3:42 PM, Hardeep Sharma wrote:
> > On 8/17/2026 6:54 PM, Konrad Dybcio wrote:
> > > On 8/14/26 9:31 AM, Hardeep Sharma wrote:
> > > > Add the TLMM pinctrl driver for the Qualcomm Kuno platform, supporting
> > > > 110 GPIOs and their pin multiplexing, pin configuration and gpiolib
> > > > interface through the common pinctrl-msm driver.
> > > >
> > > > Signed-off-by: Hardeep Sharma <hardeep.sharma@xxxxxxxxxxxxxxxx>
> > > > ---
> > > [...]
> > >
> > > > +    [65] = PINGROUP(65, qup0_se1, qdss_cti_trig1, qdss_cti_trig1, NA, NA, NA, NA, NA, NA),
> > > > +    [66] = PINGROUP(66, qup0_se1, qdss_cti_trig1, qdss_cti_trig1, NA, NA, NA, NA, NA, NA),
> > > Function3 can't be selected because of repeating names
> > >
> > > In qdss_cti_trig1_groups, you have qdss_cti_trig1 specified twice.
> > > Perhaps there was a longer name that got truncated?
> > >
> > > Konrad
> > Thanks Konrad for review.
> >
> > It's a real qdss_cti_trig1 line - the in/out
> > mirror partner of funcsel 2:
> >
> >     GPIO65: fsel2 = qdss_cti_trig1_in_mirb, fsel3 = qdss_cti_trig1_out_mire
> >     GPIO66: fsel2 = qdss_cti_trig1_out_mirb, fsel3 = qdss_cti_trig1_in_mire
> >
> > Both map to the same name because the driver uses one coarse
> > "qdss_cti_trig1" for the in/out mirror variants.
> But these are physically separate functions, even if the mirror feature
> makes them seemingly one and the same
>
> > Listing the function on both funcsels is the normal sdx convention for
> > these mirror pairs - the in-tree drivers do it, including the same pins
> > on sdx65:
> >
> >     - sdx55 [16][17][22][23][54][55][94][95] → in pinctrl-sdx55.c, pins
> > 16, 17, 22, 23, 54, 55, 94, 95 each repeat qdss_cti at two funcsels.
> >     - sdx65 [65]/[66] → in pinctrl-sdx65.c, pins 65 and 66 do it —
> >     - sdx75 [16][52][53][56][57][78][79] → in pinctrl-sdx75.c, those pins
> > do it.
> >
> > Only funcsel 2
> > is selectable from DT (first-match), which is expected and matches those
> > drivers.
> As you described it, you found a bug (FSEL3 is not selectable), please
> fix it in all the instances you've mentioned.
>
> Konrad

Thanks for review.

Corrected in v3 : 
https://lore.kernel.org/all/20260821-kuno-soc-support-v3-0-196e2b306368@xxxxxxxxxxxxxxxx/

Best Wishes

Hardeep Sharma