[PATCH v3 2/6] pwm: mediatek: fix pwm source clock selection

[Matthias Brugger]

On 07/06/2017 08:43 AM, Zhi Mao wrote:
> On Thu, 2017-07-06 at 14:16 +0800, Zhi Mao wrote:
>> On Wed, 2017-07-05 at 13:09 +0200, Matthias Brugger wrote:
>>>
>>> On 06/30/2017 08:05 AM, Zhi Mao wrote:
>>>> In original code, the pwm output frequency is not correct
>>>> when set bit<3>=1 to PWMCON register.
>>>>
>>>> Signed-off-by: Zhi Mao <zhi.mao at mediatek.com>
>>>> ---
>>>>    drivers/pwm/pwm-mediatek.c |    2 +-
>>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>>
>>>> diff --git a/drivers/pwm/pwm-mediatek.c b/drivers/pwm/pwm-mediatek.c
>>>> index 5c11bc7..d08b5b3 100644
>>>> --- a/drivers/pwm/pwm-mediatek.c
>>>> +++ b/drivers/pwm/pwm-mediatek.c
>>>> @@ -91,7 +91,7 @@ static int mtk_pwm_config(struct pwm_chip *chip, struct pwm_device *pwm,
>>>>    	if (clkdiv > 7)
>>>>    		return -EINVAL;
>>>>    
>>>> -	mtk_pwm_writel(pc, pwm->hwpwm, PWMCON, BIT(15) | BIT(3) | clkdiv);
>>>> +	mtk_pwm_writel(pc, pwm->hwpwm, PWMCON, BIT(15) | clkdiv);
>>>
>>> Just for clarification, BIT(15) enables old PWM mode, which ignores
>>> CLKSEL (BIT(3)). Therefore setting BIT(3) does not have any effect and
>>> can be discarded.
>>>
>>> Am I correct? I took mt7623n datasheet for reference.
>>>
>>> Regards,
>>> Matthias
>>>
>> Yes, remove setting bit<3> will not take any effect.
>> PWMCON bit<3> is pwm source clock selecting register.
>> You can check the datasheet of MT7623 for details.
>>
>> Regards
>> Zhi
>>
> Hi Mattias,
> 
> Ignore the above reply, I explain this bit<3> issue for you.
> In the data sheet of MT7623:
> PWMCON bit<3> is PWM source clock selecting register
> 0: CLK=CLKSRC
> 1: CLK=CLKSRC/1625
> for example,
> bit<3>=0, PWM clk source is 26M
> bit<3>=1, PWM clk source is 16K
> The frequency of PWM output will based on this clock source
> if set bit<3>=1, it will cause the frequency of PWM output is not
> correct. I also use the oscilloscope device to measure the output,
> set bit<3>=0, the output meet expectation.
>   

Reviewed-by: Matthias Brugger <matthias.bgg at gmail.com>