Re: [PATCH 02/15] dt-bindings: clock: qcom,dispcc-sm6125: Remove unused GCC_DISP_AHB_CLK

[Marijn Suijten]

On 2023-06-24 11:08:30, Krzysztof Kozlowski wrote:
> On 24/06/2023 02:41, Marijn Suijten wrote:
> > The downsteam driver for dispcc only ever gets and puts this clock
> > without ever using it in the clocktree; this unnecessary workaround was
> > never ported to mainline, hence the driver doesn't consume this clock
> > and shouldn't be required by the bindings.
> > 
> > Fixes: 8397c9c0c26b ("dt-bindings: clock: add QCOM SM6125 display clock bindings")
> > Signed-off-by: Marijn Suijten <marijn.suijten@xxxxxxxxxxxxxx>
> > ---
> 
> In perfect would we would like to know whether hardware needs this clock
> enabled/controlled, not whether some driver needs it. I understand
> though that with lack of proper docs we rely on drivers, so:

It might only use this to figure out if those clocks have already probed
or are available.  The logic goes as follows:

	clk = devm_clk_get(&pdev->dev, "cfg_ahb_clk");
	if (IS_ERR(clk)) {
		if (PTR_ERR(clk) != -EPROBE_DEFER)
			dev_err(&pdev->dev, "Unable to get ahb clock handle\n");
		return PTR_ERR(clk);
	}
	devm_clk_put(&pdev->dev, clk);

Nothing else uses/parents cfg_ahb_clk.  Maybe with clk_ignore_unused or
similar, it gets turned on but never off again?

Indeed, a lack of documentation and comment from the manufacturer makes
it impossible to know, and ignoring it (as the driver already does)
works just fine.

> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>

Thanks!

- Marijn

> 
> Best regards,
> Krzysztof
>