Re: I2C filtering (was Re: [PATCH v2 6/9] dt-bindings: i2c: at91: add binding for enable-ana-filt)

[Eugen.Hristev]

On 27.06.2019 16:22, Alexandre Belloni wrote:
> External E-Mail
> 
> 
> On 25/06/2019 11:55:33+0200, Wolfram Sang wrote:
>> On Tue, Jun 25, 2019 at 11:31:56AM +0200, Alexandre Belloni wrote:
>>> On 25/06/2019 09:14:13+0000, Eugen.Hristev@xxxxxxxxxxxxx wrote:
>>>>> Perhaps
>>>>>
>>>>> 	microchip,digital-filter;
>>>>> 	microchip,analog-filter;
>>>>>
>>>>> ?
>>>>
>>>> Hi Peter,
>>>>
>>>> Thanks for reviewing. The name of the property does not matter much to
>>>> me, and we have properties prefixed with vendor, and some are not.
>>>>
>>>> @Alexandre Belloni: which name you think it's best ?
>>>>
>>>
>>> I'm not sure, it depends on whether Wolfram thinks it is generic enough
>>> to be used without a vendor prefix.
>>
>> I could imagine that we design a generic property for filters. The ones
>> above make me wonder, though, because they are bool. I'd think you can
>> configure the filters in some way, too?
>>
> 
> Apart from enabling the filter there is indeed one configuration
> setting, the maximum pulse width of spikes to be suppressed by the input
> filter.

Hello,

This is a number 0 to 7 (3 bits) that represents the width of the spike 
in periph clock cycles.
I am looking to see what is PADFCFG , as it's related to the PAD analog 
filter configuration. It may be unused by the filter.

Eugen

> 
>> I never used such filtering, so I am unaware of the parameters needed /
>> suitable. Quick grepping through I2C master drivers reveals that
>> i2c-stm32f7.c also handles filters, but only with default values. Maybe
>> DT configuration would be benefitial to that driver, too?
>>
>> Adding some people to CC.
>>
> 
> 
>