Re: [PATCH v2 4/4] arm64: dts: qcom: shikra: Add support for AudioCoreCC and AudioCoreCSR nodes

On Wed, Jul 22, 2026 at 04:33:21PM +0530, Imran Shaik wrote:
> 
> 
> On 22-07-2026 11:54 am, Dmitry Baryshkov wrote:
> > On Mon, Jul 13, 2026 at 06:01:34PM +0530, Imran Shaik wrote:
> > > 
> > > 
> > > On 13-07-2026 05:44 pm, Konrad Dybcio wrote:
> > > > On 7/13/26 2:07 PM, Imran Shaik wrote:
> > > > > 
> > > > > 
> > > > > On 09-07-2026 05:30 pm, Konrad Dybcio wrote:
> > > > > > On 7/8/26 8:25 PM, Imran Shaik wrote:
> > > > > > > Add support for Audio Core Clock Controller (AudioCoreCC) and Audio Core
> > > > > > > CSR nodes on Qualcomm Shikra SoC. The Audio Core Clocks and Resets support
> > > > > > > differs across Shikra variants based on Audio subsystem enablement as
> > > > > > > follows:
> > > > > > > - CQM variant: The QAIF driver runs on HLOS, hence both clocks and resets
> > > > > > > are required to be supported on HLOS.
> > > > > > > - CQS variant: The QAIF driver runs on the Modem, and required clocks are
> > > > > > > handled on Modem, so from HLOS only resets are needed.
> > > > > > > - IQS variant: no soundwire codes, hence no clocks/resets are needed.
> > > > > > > 
> > > > > > > Signed-off-by: Imran Shaik <[email protected]>
> > > > > > > ---
> > > > > > 
> > > > > > For the sake of simplification, can the resets be enabled
> > > > > > globally by default?
> > > > > > 
> > > > > 
> > > > > By default (PoR), those resets are under HW control. For audio use cases, software desserts/asserts for switching to SW control. Hence, these cannot be enabled globally by default.
> > > > 
> > > > What does enabling the provider node have to do with changing the
> > > > state of the resources?
> > > > 
> > > 
> > > My bad, I interpreted it in wrong way.
> > > 
> > > We have IQS variant, where we don't have the soundwire codes, and hence the
> > > respective clocks/resets are not required to be supported. Due to this, I
> > > have kept the status disabled, and enabling only where it is required.
> > 
> > Would it make sense to do it differently - disable them for the IQS?
> 
> Yeah, we can keep the audiocore_csr node enabled by default and disable
> it in the IQS variant.
> 
> But Konrad suggested to keep it globally enabled and not disable in IQS
> variant for simplification since no one will consume those resets in the IQS
> variant and no functional impact.
> 
> Please suggest which approach to follow here.

Konrad's approach sounds fine to me.


-- 
With best wishes
Dmitry