Re: gpiolib and pinmux interaction

Re: gpiolib and pinmux interaction

[Fried, Ramon]

On 3/18/2019 13:41, Ramon Fried wrote:
> On 3/18/19 11:21 AM, Stefan Wahren wrote:
>> Hi Ramon,
>>
>> Am 17.03.19 um 10:09 schrieb Ramon Fried:
>>> Hi Linus, Stefan,
>>>
>>> Great patch, If you need help testing it on other HW tell me.
>> yes, please. I didn't get what kind of hardware you are using.
> Hi.
>
> custom mips based SOC with sta2x11 gpio driver & proprietary pin muxing ip allowing every pin to be muxed as gpio.
>
>> Regards
>> Stefan

Hey Linus, Stefan,

I'm not sure where are we standing with the patch.
I didn't find it merged, so I just took it and tested it. (It took my a 
while, hehe)
It's working, but not as I would expect, instead of marking the gpio as 
used,
it adds [kernel] flag. Which is OK, but I think it's clearer to also 
make the "used" column. (see below)

What do you think ?

Thanks,

Ramon.

See below:

# gpioinfo
gpiochip1 - 23 lines:
         line   0:      unnamed       unused   input  active-high
         line   1:      unnamed       unused   input  active-high
         line   2:      unnamed       unused   input  active-high
         line   3:      unnamed       unused   input  active-high
         line   4:      unnamed       unused   input  active-high
         line   5:      unnamed       unused   input  active-high
         line   6:      unnamed       unused   input  active-high
         line   7:      unnamed       unused   input  active-high
         line   8:      unnamed       unused   input  active-high [kernel]
         line   9:      unnamed       unused   input  active-high [kernel]
         line  10:      unnamed       unused   input  active-high [kernel]
         line  11:      unnamed       unused   input  active-high [kernel]
         line  12:      unnamed       unused   input  active-high
         line  13:      unnamed       unused   input  active-high

Re: gpiolib and pinmux interaction

[Fried, Ramon]

On 7/30/2019 19:54, Fried, Ramon wrote:
>
> On 3/18/2019 13:41, Ramon Fried wrote:
>> On 3/18/19 11:21 AM, Stefan Wahren wrote:
>>> Hi Ramon,
>>>
>>> Am 17.03.19 um 10:09 schrieb Ramon Fried:
>>>> Hi Linus, Stefan,
>>>>
>>>> Great patch, If you need help testing it on other HW tell me.
>>> yes, please. I didn't get what kind of hardware you are using.
>> Hi.
>>
>> custom mips based SOC with sta2x11 gpio driver & proprietary pin 
>> muxing ip allowing every pin to be muxed as gpio.
>>
>>> Regards
>>> Stefan
>
> Hey Linus, Stefan,
>
> I'm not sure where are we standing with the patch.
> I didn't find it merged, so I just took it and tested it. (It took my 
> a while, hehe)
> It's working, but not as I would expect, instead of marking the gpio 
> as used,
> it adds [kernel] flag. Which is OK, but I think it's clearer to also 
> make the "used" column. (see below)
>
> What do you think ?
>
> Thanks,
>
> Ramon.
>
> See below:
>
> # gpioinfo
> gpiochip1 - 23 lines:
>         line   0:      unnamed       unused   input  active-high
>         line   1:      unnamed       unused   input  active-high
>         line   2:      unnamed       unused   input  active-high
>         line   3:      unnamed       unused   input  active-high
>         line   4:      unnamed       unused   input  active-high
>         line   5:      unnamed       unused   input  active-high
>         line   6:      unnamed       unused   input  active-high
>         line   7:      unnamed       unused   input  active-high
>         line   8:      unnamed       unused   input  active-high [kernel]
>         line   9:      unnamed       unused   input  active-high [kernel]
>         line  10:      unnamed       unused   input  active-high [kernel]
>         line  11:      unnamed       unused   input  active-high [kernel]
>         line  12:      unnamed       unused   input  active-high
>         line  13:      unnamed       unused   input  active-high
>
Adding Bartosz.
Turns out I was running and old version of libgpiod, in current version 
it's printed "[used] instead of [kernel].
The patch from Stefan is perfectly fine from my standpoint, but the 
output of gpioinfo can be improved,

I'll post a patch to LIBGPIOD.

Thanks.
Ramon