Re: [PATCH RFC v3 3/4] mux: Add support for reading mux enable state from DT

[Peter Rosin <peda@axentia.se>]

On 2021-11-29 05:44, Aswath Govindraju wrote:
> Hi Peter,
> 
> On 25/11/21 7:22 pm, Peter Rosin wrote:
>> Hi!
>>
>> On 2021-11-23 09:12, Aswath Govindraju wrote:
>>> In some cases, we might need to provide the state of the mux to be set for
>>> the operation of a given peripheral. Therefore, pass this information using
>>> the second argument of the mux-controls property.
>>>
>>> Signed-off-by: Aswath Govindraju <a-govindraju@ti.com>
>>> ---
>>>  drivers/mux/core.c           | 146 ++++++++++++++++++++++++++++++++++-
>>>  include/linux/mux/consumer.h |  19 ++++-
>>>  include/linux/mux/driver.h   |  13 ++++
>>>  3 files changed, 173 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/drivers/mux/core.c b/drivers/mux/core.c
>>> index 22f4709768d1..9622b98f9818 100644
>>> --- a/drivers/mux/core.c
>>> +++ b/drivers/mux/core.c
>>> @@ -370,6 +370,29 @@ int mux_control_select_delay(struct mux_control *mux, unsigned int state,
>>>  }
>>>  EXPORT_SYMBOL_GPL(mux_control_select_delay);
>>>  
>>> +/**
>>> + * mux_state_select_delay() - Select the enable state in mux-state
>>
>> The terminology is that you have a "mux" with different "states" that you
>> "select". What you are referring to as enabling a mux state, is elsewhere
>> referred to as selecting the mux state.
>>
> 
> Sorry, for mentioning what I mean by enable state. So, the idea is the
> the state that would be mentioned in the DT property would be the state
> to which the mux to be set for enabling the given device and hence I am
> referring to it as enable state. I feel that referring to it as state
> would not convey the above.

Ah, but that this state it is use to "enable" your device is a mux
consumer detail in the context of the phy-can driver. Some other
driver might need a specific mux state for something completely
unrelated. So, the "enable" naming should not spread into the mux
code.

The situation is similar to when a driver needs an enable-gpio, the
gpio consumer knows that it's a gpio used to enable the device (or
whatever), but the gpio subsystem does not bother at all with what
the gpio is used for.

Cheers,
Peter