Re: [PATCH] dt: pinctrl: Document device tree binding

[Dong Aisheng]

On Sat, Mar 10, 2012 at 02:14:33AM +0800, Stephen Warren wrote:
> The core pin controller bindings define:
> * The fact that pin controllers expose pin configurations as nodes in
>   device tree.
> * That the bindings for those pin configuration nodes is defined by the
>   individual pin controller drivers.
> * A standardized set of properties for client devices to define numbered
>   or named pin configuration states, each referring to some number of the
>   afore-mentioned pin configuration nodes.
> * That the bindings for the client devices determines the set of numbered
>   or named states that must exist.
> 
> Signed-off-by: Stephen Warren <swarren@xxxxxxxxxxxxx>

It's exactly what we discussed in Linaro Connect.
Overally it looks good to me.
Only one small question below:

> ---
>  .../bindings/pinctrl/pinctrl-bindings.txt          |  118 ++++++++++++++++++++
>  1 files changed, 118 insertions(+), 0 deletions(-)
>  create mode 100644 Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt
> 
> diff --git a/Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt b/Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt
> new file mode 100644
> index 0000000..cce9f01
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt
> @@ -0,0 +1,118 @@
> +== Introduction ==
> +
> +Hardware modules that control pin multiplexing or configuration parameters
> +such as pull-up/down, tri-state, drive-strength etc are designated as pin
> +controllers. Each pin controller must be represented as a node in device tree,
> +just like any other hardware module.
> +
> +Hardware modules whose signals are affected by pin configuration are
> +designated client devices. Again, each client device must be represented as a
> +node in device tree, just like any other hardware module.
> +
> +For a client device to operate correctly, certain pin controllers must
> +set up certain specific pin configurations. Some client devices need a
> +single static pin configuration, e.g. set up during initialization. Others
> +need to reconfigure pins at run-time, for example to tri-state pins when the
> +device is inactive. Hence, each client device can define a set of named
> +states. The number and names of those states is defined by the client device's
> +own binding.
> +
> +The common pinctrl bindings defined in this file provide an infra-structure
> +for client device device tree nodes to map those state names to the pin
> +configuration used by those states.
> +
> +Note that pin controllers themselves may also be client devices of themselves.
> +For example, a pin controller may set up its own "active" state when the
> +driver loads. This would allow representing a board's static pin configuration
> +in a single place, rather than splitting it across multiple client device
> +nodes. The decision to do this or not somewhat rests with the author of
> +individual board device tree files, and any requirements imposed by the
> +bindings for the individual client devices in use by that board, i.e. whether
> +they require certain specific named states for dynamic pin configuration.
> +
> +== Pinctrl client devices ==
> +
> +For each client device individually, every pin state is assigned an integer
> +ID. These numbers start at 0, and are contiguous. For each state ID, a unique
> +property exists to define the pin configuration. Each state may also be
> +assigned a name. When names are used, another property exists to map from
> +those names to the integer IDs.
> +
> +Each client device's own binding determines the set of states the must be
> +defined in its device tree node, and whether to define the set of state
> +IDs that must be provided, or whether to define the set of state names that
> +must be provided.
> +
> +Required properties:
> +pinctrl-0:	List of phandles, each pointing at a pin configuration
> +		node. These referenced pin configuration nodes must be child
> +		nodes of the pin controller that they configure. Multiple
> +		entries may exist in this list so that multiple pin
> +		controllers may be configured, or so that a state may be built
> +		from multiple nodes for a single pin controller, each
> +		contributing part of the overall configuration. See the next
> +		section of this document for details of the format of these
> +		pin configuration nodes.
> +
> +		In some cases, it may be useful to define a state, but for it
> +		to be empty. This may be required when a common IP block is
> +		used in an SoC either without a pin controller, or where the
> +		pin controller does not affect the HW module in question. If
> +		the binding for that IP block requires certain pin states to
> +		exist, they must still be defined, but may be left empty.
> +
It looks this functions similar as the PIN_MAP_DUMMY_STATE you introduced
before to address the issues that the shared IP block may need or not need
pinctrl configuration on different platforms(correct me if wrong).

Then, there may be cases like below which may look a bit confusing
to people.
device {
	pinctrl-names = "active", "idle";
	pinctrl-0;
	pinctrl-1;
};

I'm wondering if we can let each individual driver to handle this special case?
Like checking device id then make decision whether call pinctrl_* APIs.
Then we can just do not define those properties for devices who
do not need pin configurations.

Regards
Dong Aisheng

> +Optional properties:
> +pinctrl-1:	List of phandles, each pointing at a pin configuration
> +		node within a pin controller.
> +...
> +pinctrl-n:	List of phandles, each pointing at a pin configuration
> +		node within a pin controller.
> +pinctrl-names:	The list of names to assign states. List entry 0 defines the
> +		name for integer state ID 0, list entry 1 for state ID 1, and
> +		so on.
> +
> +For example:
> +
> +	/* For a client device requiring named states */
> +	device {
> +		pinctrl-names = "active", "idle";
> +		pinctrl-0 = <&state_0_node_a>;
> +		pinctrl-1 = <&state_1_node_a &state_1_node_b>;
> +	};
> +
> +	/* For the same device if using state IDs */
> +	device {
> +		pinctrl-0 = <&state_0_node_a>;
> +		pinctrl-1 = <&state_1_node_a &state_1_node_b>;
> +	};
> +
> +== Pin controller devices ==
> +
> +Pin controller devices should contain the pin configuration nodes that client
> +devices reference.
> +
> +For example:
> +
> +	pincontroller {
> +		... /* Standard DT properties for the device itself elided */
> +
> +		state_0_node_a {
> +			...
> +		};
> +		state_1_node_a {
> +			...
> +		};
> +		state_1_node_b {
> +			...
> +		};
> +	}
> +
> +The contents of each of those pin configuration child nodes is defined
> +entirely by the binding for the individual pin controller device. There
> +exists no common standard for this content.
> +
> +The pin configuration nodes need not be direct children of the pin controller
> +device; they may be grand-children for example. Whether this is legal, and
> +whether there is any interaction between the child and intermediate parent
> +nodes, is again defined entirely by the binding for the individual pin
> +controller device.
> -- 
> 1.7.0.4
> 
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/