On Tue, Nov 12, 2013 at 12:23:46PM -0700, Stephen Warren wrote:

On 11/12/2013 03:59 AM, Mark Rutland wrote:

quoted

On Tue, Nov 12, 2013 at 10:47:57AM +0000, Alexander Shiyan wrote:

quoted

Hello.

quoted

On Tue, Nov 12, 2013 at 10:07:13AM +0000, Alexander Shiyan wrote:

quoted

This patch adds a new driver for the beeper controlled via GPIO pin.
The driver does not depend on the architecture and is positioned as
a replacement for the specific drivers that are used for this function.

Signed-off-by: Alexander Shiyan <redacted>

...

quoted

quoted

diff --git a/Documentation/devicetree/bindings/input/gpio-beeper.txt b/Documentation/devicetree/bindings/input/gpio-beeper.txt

...

quoted

quoted

+Example:
+
+beeper: input@0 {
+	compatible = "gpio-beeper";
+	reg = <0>;
+	gpios = <&gpio3 23 0>;
+};

What are the reg / unit-address for?

Just an example from "simple-bus" container.

If they have no meaning, they should go. They're unnecessary and make
things more confusing.

I'd expect the example to be:

beeper: beeper {
	compatible = "gpio-beeper";
	gpios - <&gpio3 23 0>;
};

And if we have multiple beepers, something like:

beeper0: beeper0 { ... };
beeper1: beeper1 { ... };

DT node names aren't meant to encode identity though. What we've done in
the past for nodes without a reg where multiple instances were desired
is to put them into simple-bus and add a reg, so:

beeper0: beeper@0 { reg = <0>; ... };
beeper1: beeper@1 { reg = <1>; ... };

Of course, if there's only one of them, then it could just be "beeper"
with no reg. The binding and example should probably reflect that simple
case.

So do we have an agreement on bindings? Otherwise the driver looks good
to me.

-- 
Dmitry