2022-03-09 01:10:05

by Bjorn Andersson

[permalink] [raw]
Subject: Re: [PATCH v10 7/7] pinctrl: qcom: Update clock voting as optional

On Tue 08 Mar 03:03 PST 2022, Srinivasa Rao Mandadapu wrote:

> Update bulk clock voting to optional voting as ADSP bypass platform doesn't
> need macro and decodec clocks,

Even I am not sure what "ADSP bypass platform" means, so please express
this better.

Are they optional because sc7280 typically come with ADSP based audio,
but it might not and if not then we shouldn't control those clocks?

> these are maintained as power domains and
> operated from lpass audio core cc.
>

So there are clocks, but they are exposed as power-domains? Or are you
just trying to say that the LPASS LPI pinctrl block is always in a
power-domain controlled by the audio clock-controller?

Regards,
Bjorn

> Signed-off-by: Srinivasa Rao Mandadapu <[email protected]>
> Co-developed-by: Venkata Prasad Potturu <[email protected]>
> Signed-off-by: Venkata Prasad Potturu <[email protected]>
> ---
> drivers/pinctrl/qcom/pinctrl-lpass-lpi.c | 12 +++++++++---
> drivers/pinctrl/qcom/pinctrl-lpass-lpi.h | 1 +
> drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c | 1 +
> 3 files changed, 11 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/pinctrl/qcom/pinctrl-lpass-lpi.c b/drivers/pinctrl/qcom/pinctrl-lpass-lpi.c
> index 1ab572f..c618b74 100644
> --- a/drivers/pinctrl/qcom/pinctrl-lpass-lpi.c
> +++ b/drivers/pinctrl/qcom/pinctrl-lpass-lpi.c
> @@ -407,9 +407,15 @@ int lpi_pinctrl_probe(struct platform_device *pdev)
> return dev_err_probe(dev, PTR_ERR(pctrl->slew_base),
> "Slew resource not provided\n");
>
> - ret = devm_clk_bulk_get(dev, MAX_LPI_NUM_CLKS, pctrl->clks);
> - if (ret)
> - return dev_err_probe(dev, ret, "Can't get clocks\n");
> + if (data->is_clk_optional) {
> + ret = devm_clk_bulk_get_optional(dev, MAX_LPI_NUM_CLKS, pctrl->clks);
> + if (ret)
> + return dev_err_probe(dev, ret, "Can't get clocks\n");
> + } else {
> + ret = devm_clk_bulk_get(dev, MAX_LPI_NUM_CLKS, pctrl->clks);
> + if (ret)
> + return dev_err_probe(dev, ret, "Can't get clocks\n");
> + }
>
> ret = clk_bulk_prepare_enable(MAX_LPI_NUM_CLKS, pctrl->clks);
> if (ret)
> diff --git a/drivers/pinctrl/qcom/pinctrl-lpass-lpi.h b/drivers/pinctrl/qcom/pinctrl-lpass-lpi.h
> index afbac2a..3bcede6 100644
> --- a/drivers/pinctrl/qcom/pinctrl-lpass-lpi.h
> +++ b/drivers/pinctrl/qcom/pinctrl-lpass-lpi.h
> @@ -77,6 +77,7 @@ struct lpi_pinctrl_variant_data {
> int ngroups;
> const struct lpi_function *functions;
> int nfunctions;
> + int is_clk_optional;
> };
>
> int lpi_pinctrl_probe(struct platform_device *pdev);
> diff --git a/drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c b/drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c
> index d67ff25..304d8a2 100644
> --- a/drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c
> +++ b/drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c
> @@ -142,6 +142,7 @@ static const struct lpi_pinctrl_variant_data sc7280_lpi_data = {
> .ngroups = ARRAY_SIZE(sc7280_groups),
> .functions = sc7280_functions,
> .nfunctions = ARRAY_SIZE(sc7280_functions),
> + .is_clk_optional = 1,
> };
>
> static const struct of_device_id lpi_pinctrl_of_match[] = {
> --
> 2.7.4
>


2022-03-17 06:16:11

by Srinivasa Rao Mandadapu

[permalink] [raw]
Subject: Re: [PATCH v10 7/7] pinctrl: qcom: Update clock voting as optional


On 3/8/2022 11:43 PM, Bjorn Andersson wrote:
Thanks for your time Bjorn!!!
> On Tue 08 Mar 03:03 PST 2022, Srinivasa Rao Mandadapu wrote:
>
>> Update bulk clock voting to optional voting as ADSP bypass platform doesn't
>> need macro and decodec clocks,
> Even I am not sure what "ADSP bypass platform" means, so please express
> this better.

Actually LPASS contains internal DSP for post processing and other
activities. Some vendors are

provided LPASS HW without enabling DSP. That's why, we call ADSP Bypass.

>
> Are they optional because sc7280 typically come with ADSP based audio,
> but it might not and if not then we shouldn't control those clocks?

They are GDSC switches. In ADSP based solution, HLOS intiate as clock to
DSP firmware and it

handles GDSC switch. In case of ADSP bypass solution,  kernel Clock
drivers handles them as power domains.

These GDSCs in genaral gets on if we vote for any clock which depends on
them.

>
>> these are maintained as power domains and
>> operated from lpass audio core cc.
>>
> So there are clocks, but they are exposed as power-domains? Or are you
> just trying to say that the LPASS LPI pinctrl block is always in a
> power-domain controlled by the audio clock-controller?
>
> Regards,
> Bjorn

Yes, They are actually GDSC switches, but they are controlled as power
domains in case of ADSP bypass case and

as clocks in ADSP based solutions.

>
>> Signed-off-by: Srinivasa Rao Mandadapu <[email protected]>
>> Co-developed-by: Venkata Prasad Potturu <[email protected]>
>> Signed-off-by: Venkata Prasad Potturu <[email protected]>
>> ---
>> drivers/pinctrl/qcom/pinctrl-lpass-lpi.c | 12 +++++++++---
>> drivers/pinctrl/qcom/pinctrl-lpass-lpi.h | 1 +
>> drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c | 1 +
>> 3 files changed, 11 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/pinctrl/qcom/pinctrl-lpass-lpi.c b/drivers/pinctrl/qcom/pinctrl-lpass-lpi.c
>> index 1ab572f..c618b74 100644
>> --- a/drivers/pinctrl/qcom/pinctrl-lpass-lpi.c
>> +++ b/drivers/pinctrl/qcom/pinctrl-lpass-lpi.c
>> @@ -407,9 +407,15 @@ int lpi_pinctrl_probe(struct platform_device *pdev)
>> return dev_err_probe(dev, PTR_ERR(pctrl->slew_base),
>> "Slew resource not provided\n");
>>
>> - ret = devm_clk_bulk_get(dev, MAX_LPI_NUM_CLKS, pctrl->clks);
>> - if (ret)
>> - return dev_err_probe(dev, ret, "Can't get clocks\n");
>> + if (data->is_clk_optional) {
>> + ret = devm_clk_bulk_get_optional(dev, MAX_LPI_NUM_CLKS, pctrl->clks);
>> + if (ret)
>> + return dev_err_probe(dev, ret, "Can't get clocks\n");
>> + } else {
>> + ret = devm_clk_bulk_get(dev, MAX_LPI_NUM_CLKS, pctrl->clks);
>> + if (ret)
>> + return dev_err_probe(dev, ret, "Can't get clocks\n");
>> + }
>>
>> ret = clk_bulk_prepare_enable(MAX_LPI_NUM_CLKS, pctrl->clks);
>> if (ret)
>> diff --git a/drivers/pinctrl/qcom/pinctrl-lpass-lpi.h b/drivers/pinctrl/qcom/pinctrl-lpass-lpi.h
>> index afbac2a..3bcede6 100644
>> --- a/drivers/pinctrl/qcom/pinctrl-lpass-lpi.h
>> +++ b/drivers/pinctrl/qcom/pinctrl-lpass-lpi.h
>> @@ -77,6 +77,7 @@ struct lpi_pinctrl_variant_data {
>> int ngroups;
>> const struct lpi_function *functions;
>> int nfunctions;
>> + int is_clk_optional;
>> };
>>
>> int lpi_pinctrl_probe(struct platform_device *pdev);
>> diff --git a/drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c b/drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c
>> index d67ff25..304d8a2 100644
>> --- a/drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c
>> +++ b/drivers/pinctrl/qcom/pinctrl-sc7280-lpass-lpi.c
>> @@ -142,6 +142,7 @@ static const struct lpi_pinctrl_variant_data sc7280_lpi_data = {
>> .ngroups = ARRAY_SIZE(sc7280_groups),
>> .functions = sc7280_functions,
>> .nfunctions = ARRAY_SIZE(sc7280_functions),
>> + .is_clk_optional = 1,
>> };
>>
>> static const struct of_device_id lpi_pinctrl_of_match[] = {
>> --
>> 2.7.4
>>