Re: [PATCH 2/4] clk: qcom: common: introduce qcom_cc_sync_state()

[Brian Masney]

Hi Konrad / Dmitry,

On Mon, Jun 15, 2026 at 04:51:35PM +0200, Konrad Dybcio wrote:
> On 6/15/26 4:48 PM, Brian Masney wrote:
> > On Mon, Jun 15, 2026 at 04:33:17PM +0200, Konrad Dybcio wrote:
> >> On 6/15/26 4:24 PM, Brian Masney wrote:
> >>> On Sun, Jun 07, 2026 at 01:30:03PM +0300, Dmitry Baryshkov wrote:
> >>>> On Sun, Jun 07, 2026 at 01:43:06AM -0300, Val Packett wrote:
> >>>>>
> >>>>> On 6/6/26 8:15 AM, Dmitry Baryshkov wrote:
> >>>>>> On Wed, Jun 03, 2026 at 10:21:47AM -0400, Brian Masney wrote:
> >>>>>>> Several qcom clk providers currently have a sync_state helper set to
> >>>>>>> icc_sync_state(). With an upcoming change to the clk framework, if
> >>>>>>> sync_state is not defined for the device, then the clk framework sets it
> >>>>>>> to clk_sync_state().
> >>>>>>> [..]
> >>>>>>> @@ -464,5 +466,12 @@ int qcom_cc_probe_by_index(struct platform_device *pdev, int index,
> >>>>>>>   }
> >>>>>>>   EXPORT_SYMBOL_GPL(qcom_cc_probe_by_index);
> >>>>>>> +void qcom_cc_sync_state(struct device *dev)
> >>>>>>> +{
> >>>>>>> +	icc_sync_state(dev);
> >>>>>> Only if desc->icc_hws != 0, otherwise it will mess the interconnect
> >>>>>> internals. You might need to set drvdata to desc.
> >>>>>
> >>>>> Hmm…
> >>>>>
> >>>>> Currently icc_sync_state does not seem to use the dev argument at all.
> >>>>>
> >>>>> How would something get messed up, now or whenever icc_sync_state changes?
> >>>>> o.0
> >>>>
> >>>> Yes :-(
> >>>
> >>> Sorry about the delayed response since I was out of town all last week.
> >>> Just to be clear, the missing check for 'desc->icc_hws != 0' is a bug that
> >>> existed prior to my change, and I should label it as such with a Fixes
> >>> tag when I post my next version?
> >>
> >> Up until this change, having icc_hws > 0 but lacking icc_sync_state
> >> (or the reverse) would be be considered programmer error
> > 
> > icc_hws > 0 but lacking icc_sync_state (or the reverse) makes sense as a
> > programmer error. However...
> > 
> >> Starting with patch 4, this gets assigned unconditionally, so there's
> >> no prior bug to be fixed
> > 
> > I don't see where that situation happens here. All of the places where
> > icc_sync_state() was previously called, the new code now calls
> > qcom_cc_sync_state() -> icc_sync_state(). (There is
> > qcom_msm8996_cbf_icc_sync_state() that needs to be modified.)
> > 
> > In patch 4 of this series, it sets up a framework level sync_state()
> > callback with dev_set_drv_sync_state(). If a sync_state already exists,
> > then that call will fail with -EBUSY, and it will leave the existing
> > sync_state() intact. So it's not calling sync_state twice. I will
> > clarify that on the comment.
> 
> Dmitry and I are referring to the situation where the clock driver isn't
> an interconnect provider but icc_sync_state() still executes. That could
> not have been the case before, since most clock drivers didn't come with
> any sort of .sync_state()

I'm sorry if I am being really dense here.

Let's ignore clk-cbf-8996.c since that has a separate issue. The other 6
drivers in this patch today all have this pattern:

static struct platform_driver foo_driver = {
	...
        .driver = {
		...
                .sync_state = icc_sync_state,
        },
};

I'm changing it to call qcom_cc_sync_state(), which will call
icc_sync_state(). So everywhere where icc_sync_state() is called today
it will still be called after the series is applied.

All of the other clk drivers will just call clk_sync_state() directly
that's set at the framework level.

Brian