Re: [PATCH v12 1/6] soc: qcom: ice: Add OPP-based clock scaling support for ICE
From: Abhinaba Rakshit
Date: Wed Oct 14 2026 - 07:32:08 EST
On Fri, Aug 28, 2026 at 02:41:50AM +0530, Abhinaba Rakshit wrote:
> On Thu, Aug 27, 2026 at 12:17:02PM +0530, Kuldeep Singh wrote:
> > On 24-08-2026 01:17, Abhinaba Rakshit wrote:
> > > Register optional operation-points-v2 table for ICE device
> > > during device probe. Attach the OPP-table with only the ICE
> > > core clock. Since, dtbinding is on a transition phase to include
> > > iface clock and clock-names, attaching the opp-table to core clock
> > > remains optional such that it does not cause probe failures.
> >
> > iface clock and corresponding dtbindings are merged for v7.3 and your
> > change will land post that.
> > We can drop this from commit message?
> >
> > I think we should scale only core clock and for iface clock should rely
> > on ufs/eMMC controller maitained frequency.
> >
>
> The idea here is to scale only the core clock and I am achieving it by using
> dev_pm_opp_set_clkname() API. However, for that clock-names needs to be as
> required property under DT bindings for qcom ICE.
>
> As the comment mentioned here: https://lore.kernel.org/all/b32c7091-b2c4-443d-b58e-759b471f67db@xxxxxxxxxxxxxxxx/,
> the current dtbindings supports backward compatibility with iface being optional,
> hence it is expected that driver should be backward compatible as well.
> Once, we plan to break this backward compatibility and make iface clock and
> clock-names property as required, we can say its out of transitional phase.
Kuldeep, I hope this answers your query.
Can you please let me know, if anything pending in this patch?
Abhinaba Rakshit