Am Dienstag, den 22.04.2014, 09:23 +0200 schrieb Thierry Reding: > On Mon, Apr 21, 2014 at 01:43:18PM -0600, Stephen Warren wrote: > > On 04/17/2014 06:02 AM, Thierry Reding wrote: > > > From: Thierry Reding <treding@xxxxxxxxxx> > > > > > > Properties referencing GPIOs should use the plural suffix -gpios. This > > > convention is encoded in the device tree backend of gpiod_get(), which > > > we'll eventually want to migrate to. > > > > Wouldn't it be simpler to fix the GPIO binding documentation and > > gpiod_get() code to allow the -gpio suffix in addition to -gpios? It > > always struck me as silly that the binding required a plural property > > name when only a single entry made sense. > > > > (For something like "clocks", since the property name applies to any > > clock, and there certainly can be many clocks, a plural property name > > makes sense. However, since each type of GPIO is "foo-gpios" rather than > > an "foo" entry in "gpios", that same argument doesn't apply, and a > > singular property name seems much more correct). > > Yeah, it's somewhat unfortunate that this is done inconsistently across > different subsystems. GPIO isn't the only exception here. Regulators use > a similar pattern. > > For consistency it'd be nice if we could get everyone to agree to one > scheme, but I suspect that by now we're far beyond that being a viable > option. > > I don't have a strong feeling either way, so if allowing both *-gpios > and *-gpio properties is what we want, then I can certainly come up with > a patch. > I agree with Stephen, allowing the singular form in the property name seems like a nicer solution (it makes for a less irritating property name), without the need to break existing DTs. Regards, Lucas -- Pengutronix e.K. | Lucas Stach | Industrial Linux Solutions | http://www.pengutronix.de/ | -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
