Messages in this thread | | | Date | Mon, 15 Jun 2009 13:50:25 +0100 | From | Ben Dooks <> | Subject | Re: [PATCH] gpiolib: Add gpio_detect, gpio_debounce and gpio_alt_func features to GPIOLIB |
| |
On Mon, Jun 15, 2009 at 11:02:53AM +0100, Mark Brown wrote: > On Mon, Jun 15, 2009 at 10:50:23AM +0100, Ben Dooks wrote: > > On Mon, Jun 15, 2009 at 05:15:02PM +0800, Alek Du wrote: > > > > * gpio_alt_func is to set pin as alternative function or GPIO. > > > gpio_alt_func is feature creep, I don't really belive this is the > > best place to put it as it will be difficult to actually make this > > generic for all gpio platforms. > > Since the proposed API just passes a value through to the driver for the > GPIO chip it looks generic enough - each chip can define whatever set of > constants it likes. I'd expect a large proportion of driver specific > APIs would end up just the same. > > Given the number of manufacturers that don't use a separate term like > the PXA MFP for the alternative functions of their GPIOs it makes sense > to have a gpiolib API for this. Without one you end up having each > driver needing to add its own API, and since the pins are just referred > to as GPIOs in the documentation the API will have that in the name and > look like it ought to be connected with gpiolib.
Yes, however I can see some horrible problems ahead as soon as people try and then try and standardise the values passed through this. The GPIO API was meant to be a lightweight way of allowing drivers at GPIOs, now everyone seems to want to push whatever they feel like in.
-- Ben (ben@fluff.org, http://www.fluff.org/)
'a smiley only costs 4 bytes'
| |