Messages in this thread | | | Date | Mon, 9 Dec 2013 11:14:21 +0000 | From | Mark Brown <> | Subject | Re: [rtc-linux] [PATCH v3 1/2] dt-bindings: add hym8563 binding |
| |
On Sat, Dec 07, 2013 at 02:46:11PM +0100, Heiko Stübner wrote: > Am Montag, 2. Dezember 2013, 14:41:10 schrieb Mark Brown: > > On Sun, Dec 01, 2013 at 08:47:42PM +0100, Heiko Stübner wrote:
> > > +Required properties: > > > +- compatible: should be: "haoyu,hym8563" > > > +- reg: i2c address > > > +- gpios: alarm interrupt gpio
> > Why is this specified as a GPIO and not as an interrupt? > > sorry for the late reply, but it seems I got somehow droppen from your > recipient list, so just found this mail on the mailinglist.
Your mail had reply to set on it.
> In v1 I specified the interrupt and the gpio. Apart from the resulting > duplication of information this also resulted in the gpio only being requested > but never used itself, which Mark Rutland did not seem to like this much :-) . > > As I'd like to keep the sanity check that really requesting the interrupt gpio > always provided thru a gpio. As there are other drivers going this route it > looked like an ok way to go.
> So what would be the real way to go? Specify only the interrupt, only the gpio > or both?
Specify only the interrupt if it's genuinely an interrupt - requiring a GPIO is broken as not all interrupt controllers are also GPIOs. There are some OMAP drivers that are broken in this regard but they shouldn't be doing that. Only use a GPIO specifier if it's used as a GPIO. [unhandled content-type:application/pgp-signature] | |