Messages in this thread | | | From | Andy Shevchenko <> | Date | Mon, 15 Jan 2018 22:19:39 +0200 | Subject | Re: [RFC PATCH 1/2] pinctrl: add consumer variant for gpio request |
| |
On Mon, Jan 15, 2018 at 6:22 PM, Ludovic Desroches <ludovic.desroches@microchip.com> wrote:
Did I miss cover letter for this?
> Add a consumer variant to GPIO request relative functions. The goal > is to fix the bad ownership, which is arbitrary set to > "range->name:gpio", of a GPIO.
Hmm... It's supposed to be name of the owner of the pin range (pin control device name IIUC).
> There is a lack of configuration features for GPIO. For instance, > we can't set the bias. Some pin controllers manage both device's > pins and GPIOs. GPIOs can benefit from pin configuration. Usually, > a pinctrl node is used to mux the pin as a GPIO and to set up its > configuration.
Don't we have means to do that?
At least that what I see in aspeed_gpio_set_config().
Or I missed a point here?
> The pinmuxing strict mode involves that a pin which is muxed can't > be requested as a GPIO if the owner is not the same.
Any elaborated example?
> Unfortunately, > the ownership of a GPIO is set arbitrarily to "range->name:gpio". > So there is a mismatch about the ownership which prevents a device > from being the owner of the pinmuxing and requesting the same pin as > a GPIO.
> Adding some consumer variants for GPIO request stuff will allow to > pass the name of the device which requests the GPIO to not return an > error if it's also the owner of the pinmuxing.
I think we need something more generic in core than producing more API functions.
But I would like to get problem first.
> + if (consumer) > + return pin_request(pctldev, pin, consumer, range); > +
Hmm... My understanding that GPIO is just a (special) function out of pin muxing. So, doing musing is essential to get proper function out of it.
Does your hardware considers this differently? If so, I would really want to see some datasheets.
-- With Best Regards, Andy Shevchenko
| |