Quoting Taniya Das (2018-10-04 05:01:27)
>
> On 9/29/2018 12:21 AM, Stephen Boyd wrote:
> > Quoting Taniya Das (2018-09-18 03:25:38)
> >> @@ -3469,6 +3495,8 @@ enum {
> >> [GCC_QSPI_CORE_CLK_SRC] = &gcc_qspi_core_clk_src.clkr,
> >> [GCC_QSPI_CORE_CLK] = &gcc_qspi_core_clk.clkr,
> >> [GCC_QSPI_CNOC_PERIPH_AHB_CLK] = &gcc_qspi_cnoc_periph_ahb_clk.clkr,
> >> + [GCC_LPASS_Q6_AXI_CLK] = NULL,
> >> + [GCC_LPASS_SWAY_CLK] = NULL,
> >> };
> >>
> >> static const struct qcom_reset_map gcc_sdm845_resets[] = {
> >> @@ -3583,6 +3611,13 @@ static int gcc_sdm845_probe(struct platform_device *pdev)
> >> if (ret)
> >> return ret;
> >>
> >> + if (of_property_read_bool(pdev->dev.of_node, "qcom,lpass-protected")) {
> >
> > Shouldn't this be negated? So that we only add the clks when lpass isn't
> > protected?
> >
>
> I was of the opinion to add the flag only when LPASS clocks are
> required. But I am fine negating it too.
It's stating that lpass clks are protected, so presumably we wouldn't
add the property on devices without the XPU configured. This means that
most configurations would have it protected and then this flag is needed
almost all the time. O well!
