Re: [PATCH 1/4] clk: qcom: clk-rcg2: Fail Duty-Cycle configuration if MND divider is not enabled.

[Nikita Travkin <nikita@xxxxxxx>]

Hi,

Stephen Boyd писал(а) 08.01.2022 05:52:
> Quoting Nikita Travkin (2021-12-09 08:37:17)
>> In cases when MND is not enabled (e.g. when only Half Integer Divider is
>> used), setting D registers makes no effect. Fail instead of making
>> ineffective write.
>>
>> Fixes: 7f891faf596e ("clk: qcom: clk-rcg2: Add support for duty-cycle for RCG")
>> Signed-off-by: Nikita Travkin <nikita@xxxxxxx>
>> ---
>>  drivers/clk/qcom/clk-rcg2.c | 7 ++++++-
>>  1 file changed, 6 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/clk/qcom/clk-rcg2.c b/drivers/clk/qcom/clk-rcg2.c
>> index e1b1b426fae4..6964cf914b60 100644
>> --- a/drivers/clk/qcom/clk-rcg2.c
>> +++ b/drivers/clk/qcom/clk-rcg2.c
>> @@ -396,7 +396,7 @@ static int clk_rcg2_get_duty_cycle(struct clk_hw *hw, struct clk_duty *duty)
>>  static int clk_rcg2_set_duty_cycle(struct clk_hw *hw, struct clk_duty *duty)
>>  {
>>         struct clk_rcg2 *rcg = to_clk_rcg2(hw);
>> -       u32 notn_m, n, m, d, not2d, mask, duty_per;
>> +       u32 notn_m, n, m, d, not2d, mask, duty_per, cfg;
>>         int ret;
>>
>>         /* Duty-cycle cannot be modified for non-MND RCGs */
>> @@ -407,6 +407,11 @@ static int clk_rcg2_set_duty_cycle(struct clk_hw *hw, struct clk_duty *duty)
>>
>>         regmap_read(rcg->clkr.regmap, RCG_N_OFFSET(rcg), &notn_m);
>>         regmap_read(rcg->clkr.regmap, RCG_M_OFFSET(rcg), &m);
>> +       regmap_read(rcg->clkr.regmap, RCG_CFG_OFFSET(rcg), &cfg);
>> +
>> +       /* Duty-cycle cannot be modified if MND divider is in bypass mode. */
>> +       if (!(cfg & CFG_MODE_MASK))
>> +               return -EINVAL;
> 
> Should we still allow 50% duty cycle to succeed?

*Technically* setting 50% duty cycle works since it's the default,
but how I understand it, the main way to get there is to call
clk_set_duty_cycle() which implies that it's caller intends
to control duty cycle specifically but the call doesn't actually
control anything as the hardware block is disabled.

I'm adding this error here primarily to bring attention of the
user (e.g. developer enabling some peripheral that needs
duty cycle control) who might have to change their clock tree
to make this control effective. So, assuming that if someone
sets the duty cycle to 50% then they might set it to some other
value later, it makes sense to fail the first call anyway.

If you think there are some other possibilities for this call
to happen specifically with 50% duty cycle (e.g. some
preparations or cleanups in the clk subsystem or some drivers
that I'm not aware of) then I can make an exemption in the check
for that.

Thanks,
Nikita