[RFC v2] dt-binding: media: document ON Semi AR0521 sensor bindings

[RFC v2] dt-binding: media: document ON Semi AR0521 sensor bindings

[Krzysztof Hałasa]

This file documents DT bindings for the AR0521 camera sensor driver.
Changes from v1:
- added power management (power supplies).
- small fixes

The question still stands: is there a way to reliably put national
unicode characters into:
- commit messages for patches submitted via email,
- C and other source files (comments and stuff like MODULE_AUTHOR).

Yes, I know I can commit it myself correctly, but then propagating it
upstream is problematic. Perhaps a pullable tree would be better?
I guess I need to renew my old kernel.org account.

Signed-off-by: Krzysztof Halasa <khalasa@piap.pl>

diff --git a/Documentation/devicetree/bindings/media/i2c/onnn,ar0521.yaml b/Documentation/devicetree/bindings/media/i2c/onnn,ar0521.yaml
new file mode 100644
index 000000000000..29421daacc87
--- /dev/null
+++ b/Documentation/devicetree/bindings/media/i2c/onnn,ar0521.yaml
@@ -0,0 +1,87 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/media/i2c/onnn,ar0521.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: ON Semiconductor AR0521 MIPI CSI-2 sensor
+
+maintainers:
+  - Krzysztof Halasa <khalasa@piap.pl>
+
+description: |-
+  The AR0521 is a raw CMOS image sensor with MIPI CSI-2 and
+  I2C-compatible control interface.
+
+properties:
+  compatible:
+    const: onnn,ar0521
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    description: reference to the xclk clock
+    maxItems: 1
+
+  clock-names:
+    const: xclk
+
+  vdd_io-supply:
+    description:
+      Definition of the regulator used as digital I/O (1.8 V) voltage supply.
+
+  vdd_core-supply:
+    description:
+      Definition of the regulator used as digital core (1.2 V) voltage supply.
+
+  vcc_analog-supply:
+    description:
+      Definition of the regulator used as analog (2.7 V) voltage supply.
+
+  reset-gpios:
+    description: reset GPIO, usually active low
+    maxItems: 1
+
+  port:
+    $ref: /schemas/graph.yaml#/properties/port
+    description: |
+      Output video port: 1, 2 or 4 lanes.
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - clock-names
+  - port
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+    #include <dt-bindings/clock/imx6qdl-clock.h>
+
+    i2c {
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            ar0521: camera-sensor@36 {
+                    compatible = "onnn,ar0521";
+                    reg = <0x36>;
+                    pinctrl-names = "default";
+                    pinctrl-0 = <&pinctrl_mipi_camera>;
+
+                    clocks = <&clks IMX6QDL_CLK_CKO>;
+                    clock-names = "xclk";
+
+                    reset-gpios = <&gpio1 7 GPIO_ACTIVE_LOW>;
+
+                    port {
+                           mipi_camera_to_mipi_csi2: endpoint {
+                                    remote-endpoint = <&mipi_csi2_in>;
+                                    data-lanes = <1 2 3 4>;
+                            };
+                    };
+            };
+    };

-- 
Krzysztof Hałasa

Sieć Badawcza Łukasiewicz
Przemysłowy Instytut Automatyki i Pomiarów PIAP
Al. Jerozolimskie 202, 02-486 Warszawa

Re: [RFC v2] dt-binding: media: document ON Semi AR0521 sensor bindings

[Laurent Pinchart]

Hi Krzysztof,

Thank you for the patch.

On Tue, Jun 22, 2021 at 01:18:35PM +0200, Krzysztof Hałasa wrote:
> This file documents DT bindings for the AR0521 camera sensor driver.
> Changes from v1:
> - added power management (power supplies).
> - small fixes

Please move the text below after a --- line, it shouldn't be included in
the commit message.

> The question still stands: is there a way to reliably put national
> unicode characters into:
> - commit messages for patches submitted via email,

This shouldn't be too much of a problem, as long as you MUA and MTA
don't mess up encoding.

> - C and other source files (comments and stuff like MODULE_AUTHOR).

This may be more problematic, especially in strings in source code.

> Yes, I know I can commit it myself correctly, but then propagating it
> upstream is problematic. Perhaps a pullable tree would be better?
> I guess I need to renew my old kernel.org account.
> 
> Signed-off-by: Krzysztof Halasa <khalasa@piap.pl>
> 
> diff --git a/Documentation/devicetree/bindings/media/i2c/onnn,ar0521.yaml b/Documentation/devicetree/bindings/media/i2c/onnn,ar0521.yaml
> new file mode 100644
> index 000000000000..29421daacc87
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/media/i2c/onnn,ar0521.yaml
> @@ -0,0 +1,87 @@
> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/media/i2c/onnn,ar0521.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: ON Semiconductor AR0521 MIPI CSI-2 sensor
> +
> +maintainers:
> +  - Krzysztof Halasa <khalasa@piap.pl>
> +
> +description: |-
> +  The AR0521 is a raw CMOS image sensor with MIPI CSI-2 and
> +  I2C-compatible control interface.
> +
> +properties:
> +  compatible:
> +    const: onnn,ar0521
> +
> +  reg:
> +    maxItems: 1
> +
> +  clocks:
> +    description: reference to the xclk clock

You can drop the description, it's implied by the clock-names property.

> +    maxItems: 1
> +
> +  clock-names:
> +    const: xclk

Isn't the pin named extclk ?

> +
> +  vdd_io-supply:
> +    description:
> +      Definition of the regulator used as digital I/O (1.8 V) voltage supply.
> +
> +  vdd_core-supply:
> +    description:
> +      Definition of the regulator used as digital core (1.2 V) voltage supply.

It's not just the digital core, 1.2V is also needed for the PLL
(VDD_PLL) and the PHY (VDD_PHY). I'd call this vdd-supply.

> +  vcc_analog-supply:
> +    description:
> +      Definition of the regulator used as analog (2.7 V) voltage supply.

Similarly, I'd call this vaa-supply is is covers VAA and VAA_PIX
(there's no VCC_ANALOG pin).

> +
> +  reset-gpios:
> +    description: reset GPIO, usually active low
> +    maxItems: 1
> +
> +  port:
> +    $ref: /schemas/graph.yaml#/properties/port
> +    description: |
> +      Output video port: 1, 2 or 4 lanes.

The number of lanes should be described through the data-lanes endpoint
property.

> +
> +required:
> +  - compatible
> +  - reg
> +  - clocks
> +  - clock-names
> +  - port

Supplies should be mandatory as the chip needs them.

> +
> +additionalProperties: false
> +
> +examples:
> +  - |
> +    #include <dt-bindings/gpio/gpio.h>
> +    #include <dt-bindings/clock/imx6qdl-clock.h>
> +
> +    i2c {
> +            #address-cells = <1>;
> +            #size-cells = <0>;
> +
> +            ar0521: camera-sensor@36 {
> +                    compatible = "onnn,ar0521";
> +                    reg = <0x36>;
> +                    pinctrl-names = "default";
> +                    pinctrl-0 = <&pinctrl_mipi_camera>;
> +
> +                    clocks = <&clks IMX6QDL_CLK_CKO>;
> +                    clock-names = "xclk";
> +
> +                    reset-gpios = <&gpio1 7 GPIO_ACTIVE_LOW>;
> +
> +                    port {
> +                           mipi_camera_to_mipi_csi2: endpoint {
> +                                    remote-endpoint = <&mipi_csi2_in>;
> +                                    data-lanes = <1 2 3 4>;
> +                            };
> +                    };
> +            };
> +    };

-- 
Regards,

Laurent Pinchart

Re: [RFC v2] dt-binding: media: document ON Semi AR0521 sensor bindings

[Krzysztof Hałasa]

Laurent Pinchart <laurent.pinchart@ideasonboard.com> writes:

>> The question still stands: is there a way to reliably put national
>> unicode characters into:
>> - commit messages for patches submitted via email,
>
> This shouldn't be too much of a problem, as long as you MUA and MTA
> don't mess up encoding.

Maybe it's better now. I had mixed results in the past, but maybe it was
10+ years ago. Then I stopped using non-ASCII as they weren't very
essential.
Apparently there was no such problems with drivers/net, at least from
the time I started using non-ASCII characters.
-- 
Krzysztof Hałasa

Sieć Badawcza Łukasiewicz
Przemysłowy Instytut Automatyki i Pomiarów PIAP
Al. Jerozolimskie 202, 02-486 Warszawa

Re: [RFC v2] dt-binding: media: document ON Semi AR0521 sensor bindings

[Laurent Pinchart]

On Wed, Jun 23, 2021 at 07:46:25AM +0200, Krzysztof Hałasa wrote:
> Laurent Pinchart <laurent.pinchart@ideasonboard.com> writes:
> 
> >> The question still stands: is there a way to reliably put national
> >> unicode characters into:
> >> - commit messages for patches submitted via email,
> >
> > This shouldn't be too much of a problem, as long as you MUA and MTA
> > don't mess up encoding.
> 
> Maybe it's better now. I had mixed results in the past, but maybe it was
> 10+ years ago. Then I stopped using non-ASCII as they weren't very
> essential.
> Apparently there was no such problems with drivers/net, at least from
> the time I started using non-ASCII characters.

I think it should be better, yes. As long as there's no MS Exchange
along the way of course ;-)

-- 
Regards,

Laurent Pinchart