linux-kernel - Re: [PATCH v2] regulator: gpio-regulator: add DT bindings

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <504E4234.9090300@wwwdotorg.org>
Date:	Mon, 10 Sep 2012 13:40:36 -0600
From:	Stephen Warren <swarren@...dotorg.org>
To:	Daniel Mack <zonque@...il.com>
CC:	devicetree-discuss@...ts.ozlabs.org, lrg@...com,
	broonie@...nsource.wolfsonmicro.com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] regulator: gpio-regulator: add DT bindings

On 09/10/2012 09:03 AM, Daniel Mack wrote:
> Add DT bindings for the gpio-regulator driver and some documentation on
> how to use it.

> +++ b/Documentation/devicetree/bindings/regulator/gpio-regulator.txt
> @@ -0,0 +1,58 @@
> +Device tree bindings for GPIO controlled voltages
> +
> +Voltage or current regulators on boards that are controlled by GPIOs can
> +be used by the gpio-regulator driver.
> +
> +Required properties:
> +
> + - "compatible":	must be set to "gpio-regulator"
> + - "regulator-name": 	must be set to a string describing the name of this
> +			regulator.

I'd expect the "gpios" property to be listed here.

> +Optional properties:
> +
> + - "startup-delay":		Start-up time in microseconds
> + - "enable-high":		Polarity of enable GPIO. Active high if
> +				defined, otherwise active low

fixed-regulator.txt already has an "enable-active-high" property; it'd
be nice to be consistent.

> + - "enabled-at-boot":		If set, the regulator has been enabled at boot
> +				time

Isn't that regulator-boot-on, as defined in regulator.txt that you
mentioned?

> + - "regulator-type-voltage":	The regulator controls a voltage
> + - "regulator-type-current":	The regulator controls a current

I wonder if it'd be better to differentiate this using different
compatible values instead?

> + - "states":			An array of possible states, describing the
> +				GPIO states to reach a given value
> +	- "value":		The value of the state, in microvolts or
> +				microamperes

A name like "microvolts", "voltage", "microamps", or "current" might be
more descriptive here.

> +	- "gpios":		bitfield of gpio target-states for the value
> +				The n-th bit in the bitfield describes the
> +				state of the n-th GPIO from the gpios-array

"gpios" sounds like the name of a property that defines which GPIOs are
used, rather than the value of those GPIOs. Perhaps "gpio-values" instead?

Actually, I wonder if we really even need to represent that values
explicitly; rather than have a big set of nodes, perhaps you can instead
have a single property that lists the voltage (or current) that each
combination of GPIO values gives; something like:

reg_gpio {
    ...
    gpios = <&ggpio 23 0 &gpio 24 0>;
    voltages = <0 1000000 2000000 3000000>;
};

(As inspiration for that, I looked at
Documentation/devicetree/bindings/net/mdio-mux-gpio.txt which doesn't
have a property defining the value of the GPIOs for each combination,
but rather a node per valid combination).

Perhaps represent and invalid GPIO combination with 0xffffffff?

> +Also, all properties described in
> +Documentation/devicetree/bindings/regulator/regulator.txt are supported as
> +well.
> +
> +Example:
> +
> +	reg_gpio {
> +		compatible = "gpio-regulator";
> +		regulator-name = "voltage";
> +		regulator-enable-high;
> +		regulator-type-voltage;
> +		regulator-min-microvolt = <3300000>;
> +		regulator-max-microvolt = <3300000>;
> +		regulator-always-on;
> +
> +		gpios = <&gpio 23 0>;
> +
> +		states {
> +			state-on {
> +				value = <3300000>;
> +				gpios = <0x1>;
> +			};
> +
> +			state-off {
> +				value = <0>;
> +				gpios = <0x0>;
> +			};
> +		};
> +	};

That particular example only has on/off states, and so might be better
covered using the existing fixed-regulator, with optional GPIO control.
Perhaps an example using 2 GPIOs for 4 voltage states would be more useful?
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/