[PATCH 3/4] arm64: dts: qcom: Add Shikra CQM SoM platform

[Dmitry Baryshkov]

On Fri, Jul 03, 2026 at 03:00:35PM +0530, Rakesh Kota wrote:
> On Tue, Jun 30, 2026 at 03:50:16PM +0200, Konrad Dybcio wrote:
> > On 6/30/26 2:42 PM, Rakesh Kota wrote:
> > > On Sun, Jun 28, 2026 at 03:33:23PM +0300, Dmitry Baryshkov wrote:
> > >> On Thu, Jun 25, 2026 at 09:11:19PM +0530, Kamal Wadhwa wrote:
> > >>> On Wed, Jun 17, 2026 at 03:48:14PM +0300, Dmitry Baryshkov wrote:
> > >>>> On Mon, 18 May 2026 at 14:49, Kamal Wadhwa
> > >>>> <kamal.wadhwa at oss.qualcomm.com> wrote:
> > >>>>>
> > >>>>> On Sun, May 17, 2026 at 08:18:15PM +0300, Dmitry Baryshkov wrote:
> > >>>>>> On Thu, May 14, 2026 at 04:09:18PM +0530, Kamal Wadhwa wrote:
> > >>>>>>> On Wed, May 13, 2026 at 06:14:20PM +0300, Dmitry Baryshkov wrote:
> > >>>>>>>> On 13/05/2026 17:29, Rakesh Kota wrote:
> > >>>>>>>>> On Wed, May 13, 2026 at 03:01:47PM +0300, Dmitry Baryshkov wrote:
> > >>>>>>>>>> On Wed, May 13, 2026 at 04:28:35AM +0000, sashiko-bot at kernel.org wrote:
> > >>>>>>>>>>> Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
> > >>>>>>>>>>> - [High] The PMIC regulator definitions omit their required input supply dependencies (e.g., `vdd_s2-supply`, `vdd_l3-supply`), breaking the power hierarchy.
> > >>>>>>>>>>> - [Medium] The device tree inaccurately hardcodes the `compatible` string to a different PMIC model (`qcom,rpm-pm2250-regulators`) instead of explicitly identifying the actual hardware (PM4125).
> > >>>>>>>>>>> --
> > >>>>>>>>>>>> +
> > >>>>>>>>>>>> +         pm4125_s2: s2 {
> > >>>>>>>>>>>> +                 regulator-min-microvolt = <1000000>;
> > >>>>>>>>>>>> +                 regulator-max-microvolt = <1200000>;
> > >>>>>>>>>>>> +         };
> > >>>>>>>>>>>
> > >>>>>>>>>>> Do these regulators need to explicitly define their input supply dependencies
> > >>>>>>>>>>> such as vdd_s2-supply?
> > >>>>>>>>>>>
> > >>>>>>>>>>> Without these properties, the regulator framework might be unaware that the
> > >>>>>>>>>>> PMIC regulators draw power from upstream supplies.
> > >>>>>>>>>>>
> > >>>>>>>>>>> If the kernel dynamically manages the upstream supply and its reference count
> > >>>>>>>>>>> drops to zero, could it be disabled, causing an unexpected power loss for
> > >>>>>>>>>>> downstream components?
> > >>>>>>>>>>
> > >>>>>>>>>> And this is a correct comment. Please provide missing supplies.
> > >>>>>>>>>>
> > >>>>>>>>> As per the Qualcomm system design, the parent-child supply relationship
> > >>>>>>>>> is managed by the RPM firmware, not the Linux regulator framework. The
> > >>>>>>>>> RPM ensures the parent supply is never disabled until all subsystem
> > >>>>>>>>> votes are cleared.
> > >>>>>>>>
> > >>>>>>>> How is this different from other, previous platforms?
> > >>>>>>>
> > >>>>>>> This is not different. In the previous platforms too this is taken care from the
> > >>>>>>> RPM/RPMH firmware side, the only case where we may need explicit vote to parent
> > >>>>>>> is for non-rpmh/rpm regulator rails (like i2c based regulator pm8008), which
> > >>>>>>> may have a RPM/RPMH regulator as a parent.
> > >>>>>>>
> > >>>>>>> Even on those previous targets the parent rail of all RPM/RPMH regulators are
> > >>>>>>> internally voted by RPM/RPMH FW at proper voltage with required headroom
> > >>>>>>> calculated based on the active child rails. This was done for all the
> > >>>>>>> subsystems (including APPS) regulators.
> > >>>>>>>
> > >>>>>>> So no explicit handling from the APPS is required for parent supply.
> > >>>>>>
> > >>>>>> You are explaining the driver behaviour. But the question is about the
> > >>>>>> hardware description. If there is no difference, please add necessary
> > >>>>>> supplies back.
> > >>>>>
> > >>>>> I understand your concern about descibing the parent-child relation in the
> > >>>>> devicetree, and given that we have been almost always followed this for all
> > >>>>> the previous targets, it will expected of us to add them.
> > >>>>
> > >>>> Yes.
> > >>>>
> > >>>>>
> > >>>>> However, we want to avoid the unnecessary access to the parent from APPS.
> > >>>>
> > >>>> Why? What is the reason? Do we want to do the same for all the
> > >>>> platforms? Only for Shikra? Something else?
> > >>>>
> > >>>>> At the moment, I do not see a way to avoid that, if we add the parent
> > >>>>> regulators.
> > >>>>
> > >>>> That depend on the answer to the previous question. In the end, we can
> > >>>> make the driver ignore the parents by removing them from the regulator
> > >>>> desc.
> > >>>
> > >>> Ok, this seems like a good suggestion, so you mean its ok if we define the
> > >>> regulator desc's supply column with NULL? And only keep that in the DT?
> > >>>
> > >>> you mean like this?
> > >>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/regulator/qcom-rpmh-regulator.c?h=v7.1#n1453
> > >>>
> > >>> (please let me know if i got that right. thanks)
> > >>
> > >> Yes. Don't forget to explain in the commit message, why you are doing
> > >> so.
> > > 
> > > Currently, Agatti uses the same PMIC, so we cannot set the driver
> > > supply name reference to NULL. Since it's an older target,
> > > we'll need to run a regression before making any driver-level changes.
> > > 
> > > Additionally, the child-to-parent regulator ganging differs between
> > > Shikra and Agatti:
> > > 
> > >  - On Agatti, l3 regulator is ganged with vdd_l13_l14_l15_l16
> > >  - On Shikra, l3 is ganged with vdd_l2_l3
> > 
> > Is it configurable on the PMIC level? I was under the impression the
> > supply maps are fixed in hardware.  Is there a chance the agatti
> > description is just wrong?
> 
> The supply ganging between child LDOs and parent supplies is not fixed
> at the PMIC hardware level — it varies per platform based on system
> design requirements. The same LDO can have a different parent supply on
> different SoCs.

It can have different power supplies, but the supply is provided through
a fixed pin of the PMIC. Please correct me if I'm wrong.

> I have verified the Agatti parent-child supply mappings and they are
> correct. The difference between Agatti and Shikra is a legitimate
> platform-level design difference, not an error in the Agatti DT.

There is an error in the DT. It uses fake PMIC supply names.

-- 
With best wishes
Dmitry