Re: [PATCH v7 2/2] cpufreq: qcom-hw: Add support for QCOM cpufreq HW driver

[Stephen Boyd]

Quoting Taniya Das (2018-08-08 03:15:26)
> 
> 
> On 8/8/2018 11:52 AM, Stephen Boyd wrote:
> >>
> >> Binding describes hardware controllable by the OS. That's the reality.
> >> Let's not add mandatory clock bindings for clocks that the OS can't do
> >> anything about.
> >>
> > 
> > It seems that you believe clks should only be used to turn on/off and
> > control rates. That is not the whole truth. Sometimes clks are there
> > just to express the clk frequencies that are present in the design so
> > that drivers can figure out what to do.
> > 
> 
> Stephen,
> 
> As this clock is not configurable by linux clock drivers and we really 
> do not care the parent src(as mentioned by Saravana) to generate the 
> 300MHz, would it be good to define a fixed rate clock so as to express 
> the HW connectivity & frequency?
> 

As a hack that works great, but why do we need to workaround problems by
adding a fixed rate clk to DT for this PLL? The PLL is provided by GCC
node so it should be connected to the GCC node.