Re: [PATCH v2 1/2] media: dt-bindings: alvium: add document YAML binding

From: Laurent Pinchart
Date: Wed May 31 2023 - 07:06:15 EST

Next message: David Howells: "[PATCH net-next v2 2/2] kcm: Convert kcm_sendpage() to use MSG_SPLICE_PAGES"
Previous message: David Howells: "[PATCH net-next v2 0/2] splice, net: Handle MSG_SPLICE_PAGES in AF_KCM"
In reply to: Tommaso Merciai: "Re: [PATCH v2 1/2] media: dt-bindings: alvium: add document YAML binding"
Next in thread: Tommaso Merciai: "Re: [PATCH v2 1/2] media: dt-bindings: alvium: add document YAML binding"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hi Tommaso,

On Wed, May 31, 2023 at 12:20:47PM +0200, Tommaso Merciai wrote:
> On Mon, May 29, 2023 at 09:43:26AM +0300, Laurent Pinchart wrote:
> > On Mon, May 29, 2023 at 09:39:13AM +0300, Laurent Pinchart wrote:
> > > On Sun, May 28, 2023 at 09:16:05PM +0000, Sakari Ailus wrote:
> > > > On Fri, May 26, 2023 at 07:39:43PM +0200, Tommaso Merciai wrote:
> > > > > Add documentation of device tree in YAML schema for the ALVIUM
> > > > > Camera from Allied Vision Inc.
> > > > >
> > > > > References:
> > > > > - https://www.alliedvision.com/en/products/embedded-vision-solutions
> > > > >
> > > > > Signed-off-by: Tommaso Merciai <tomm.merciai@xxxxxxxxx>
> > > > > ---
> > > > > Changes since v1:
> > > > > - Fixed build error as suggested by RHerring bot
> > > > >
> > > > > .../media/i2c/alliedvision,alvium.yaml | 115 ++++++++++++++++++
> > > > > 1 file changed, 115 insertions(+)
> > > > > create mode 100644 Documentation/devicetree/bindings/media/i2c/alliedvision,alvium.yaml
> > > > >
> > > > > diff --git a/Documentation/devicetree/bindings/media/i2c/alliedvision,alvium.yaml b/Documentation/devicetree/bindings/media/i2c/alliedvision,alvium.yaml
> > > > > new file mode 100644
> > > > > index 000000000000..81e9e560c99d
> > > > > --- /dev/null
> > > > > +++ b/Documentation/devicetree/bindings/media/i2c/alliedvision,alvium.yaml
> > > > > @@ -0,0 +1,115 @@
> > > > > +# SPDX-License-Identifier: GPL-2.0
> > > > > +%YAML 1.2
> > > > > +---
> > > > > +$id: http://devicetree.org/schemas/media/i2c/alliedvision,alvium.yaml#
> > > > > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > > > > +
> > > > > +title: Alliedvision Alvium Camera
> > >
> > > s/Alliedvision/Allied Vision/
> > >
> > > > > +
> > > > > +maintainers:
> > > > > + - Tommaso Merciai <tomm.merciai@xxxxxxxxx>
> > > > > + - Martin Hecht <martin.hecht@xxxxxxxx>
> > > > > +
> > > > > +allOf:
> > > > > + - $ref: /schemas/media/video-interface-devices.yaml#
> > > > > +
> > > > > +properties:
> > > > > + compatible:
> > > > > + const: alliedvision,alvium
> > >
> > > The name is very generic. There are Alvium camera modules that have a
> > > GMSL or FPD-Link interface, and I'm pretty sure those will require a
> > > different driver. I would add module-specific compatible strings (e.g.
> > > "alliedvision,alvium-1500c", ...) here, with a generic fallback.
> > > "alliedvision,alvium" isn't good as it won't cover GMSL or FPD-Link,
> > > maybe "alliedvision,alvium-csi2" would be an option.
> >
> > Actually, "alvium-1500c" as a specific compatible string won't do. You
> > need the exact model in the compatible string, otherwise it won't be
> > possible for the driver to handle device-specific configuration (for
> > instance accessing registers of the camera sensor for fine-grained
> > configuration). I would thus recommend using "alliedvision,alvium-1500c"
> > and "alliedvision,alvium-1800c" as generic fallbacks, along compatible
> > strings that include the exact device model.
>
> Agree with alliedvision,alvium-csi2 and thanks for your suggestion.
> In my opinion we don’t need names for 1500c and
> others because the same driver can drive all the alvium models.
> Alvium is taking care of different sensor abstractions.
>
> I test with this driver with the following models:
>
> - 1800 C-1240c
> - 1800 C-040c
> - 1500 C-500
>
> What do you think about?

As far as I understand, the camera modules allow accessing sensors
registers from the host (through the ISP) for fine-grained
configuration. Even if that's not implemented in the driver at the
moment, I think it's an important feature to eventually support, and it
will require a way for the system to identify the camera module
precisely, to know which sensor the module uses. That's why I would like
that information to be available in DT, in the form of a compatible
string. For instance,

compatible = "alliedvision,alvium-1500c-1240c",
"alliedvision,alvium-1500c";

The driver will only need DT match entries for
"alliedvision,alvium-1500c" and "alliedvision,alvium-1800c".

> > > > > +
> > > > > + reg:
> > > > > + maxItems: 1
> > > > > +
> > > > > + clocks:
> > > > > + description: XCLK Input Clock
> > > > > +
> > > > > + clock-names:
> > > > > + const: xclk
> > > >
> > > > I'd also drop this as you have a single clock only: it's redundant.
> > > >
> > > > > +
> > > > > + powerdown-gpios:
> > > > > + maxItems: 1
> > > > > + description: >
> > > > > + Reference to the GPIO connected to the powerdown pin, if any.
> > > > > +
> > > > > + reset-gpios:
> > > > > + maxItems: 1
> > > > > + description: >
> > > > > + Reference to the GPIO connected to the reset pin, if any.
> > >
> > > Reading the Alvium CSI-2 Cameras User Guide, I don't see any powerdown
> > > or reset pin on the 22-pin connector. Am I missing something ? There are
> > > however two GPIOs (in addition to the I2C signals that are also
> > > documented as GPIOs), do you plan to support those ?
> > >
> > > > > +
> > > > > + streamon-delay:
> > > > > + maxItems: 1
> > > > > + description: >
> > > > > + Delay before camera start capturing frames in us.
> > >
> > > Add "-us" to the property name to indicate the unit.
> > >
> > > This is a vendor-specific property, and should thus have a vendor
> > > prefix.
> > >
> > > A longer description is needed, from that single line I have no idea
> > > what the property does exactly.
> > >
> > > > > +
> > > > > + rotation:
> > > > > + enum:
> > > > > + - 0
> > > > > + - 180
> > >
> > > Why is the rotation restricted to 0 or 180 ? Someone could mount the
> > > module with 90 degrees rotation, shouldn't the DT bindings allow
> > > describing that ?
> > >
> > > You need a property for the vcc-ext-in supply.
> > >
> > > > > +
> > > > > + port:
> > > > > + description: Digital Output Port
> > > > > + $ref: /schemas/graph.yaml#/$defs/port-base
> > > > > + additionalProperties: false
> > > > > +
> > > > > + properties:
> > > > > + endpoint:
> > > > > + $ref: /schemas/media/video-interfaces.yaml#
> > > > > + unevaluatedProperties: false
> > > > > +
> > > > > + properties:
> > > > > + clock-lanes:
> > > > > + const: 0
> > > >
> > > > The driver can know this, no need to have it in DT, i.e. please drop it.
> > > >
> > > > > + data-lanes:
> > > > > + minItems: 1
> > > > > + maxItems: 4
> > > > > + link-frequencies: true
> > > > > +
> > > > > + required:
> > > > > + - data-lanes
> > > > > + - link-frequencies
> > > > > +
> > > > > +required:
> > > > > + - compatible
> > > > > + - reg
> > > > > + - clocks
> > > > > + - clock-names
> > > > > + - port
> > > > > +
> > > > > +additionalProperties: false
> > > > > +
> > > > > +examples:
> > > > > + - |
> > > > > + #include <dt-bindings/gpio/gpio.h>
> > > > > + #include <dt-bindings/clock/imx8mp-clock.h>
> > > > > +
> > > > > + i2c {
> > > > > + #address-cells = <1>;
> > > > > + #size-cells = <0>;
> > > > > +
> > > > > + camera: alvium@3c {
> > > > > + compatible = "alliedvision,alvium";
> > >
> > > The "alliedvision" prefix is missing from
> > > Documentation/devicetree/bindings/vendor-prefixes.yaml.
> > >
> > > > > + pinctrl-names = "default";
> > > > > + pinctrl-0 = <&pinctrl_csi0_pwn>, <&pinctrl_csi0_rst>, <&pinctrl_csi_mclk>;
> > >
> > > I'd drop pinctrl, it makes the example longer without adding much value.
> > >
> > > > > + reg = <0x3c>;
> > > > > + clocks = <&clk IMX8MP_CLK_IPP_DO_CLKO2>;
> > > > > + clock-names = "xclk";
> > > > > + assigned-clocks = <&clk IMX8MP_CLK_IPP_DO_CLKO2>;
> > > > > + assigned-clock-parents = <&clk IMX8MP_CLK_24M>;
> > > > > + assigned-clock-rates = <24000000>;
> > > > > + streamon-delay = <20>;
> > > > > + powerdown-gpios = <&gpio2 11 GPIO_ACTIVE_HIGH>;
> > > > > + reset-gpios = <&gpio1 6 GPIO_ACTIVE_LOW>;
> > > > > + status = "okay";
> > > > > +
> > > > > + port {
> > > > > + alvium_out: endpoint {
> > > > > + remote-endpoint = <&mipi_csi_0_in>;
> > > > > + data-lanes = <1 2 3 4>;
> > > > > + link-frequencies = /bits/ 64 <681250000>;
> > > > > + clock-lanes = <0>;
> > > > > + };
> > > > > + };
> > > > > + };
> > > > > + };
> > > > > +
> > > > > +...

--
Regards,

Laurent Pinchart

Next message: David Howells: "[PATCH net-next v2 2/2] kcm: Convert kcm_sendpage() to use MSG_SPLICE_PAGES"
Previous message: David Howells: "[PATCH net-next v2 0/2] splice, net: Handle MSG_SPLICE_PAGES in AF_KCM"
In reply to: Tommaso Merciai: "Re: [PATCH v2 1/2] media: dt-bindings: alvium: add document YAML binding"
Next in thread: Tommaso Merciai: "Re: [PATCH v2 1/2] media: dt-bindings: alvium: add document YAML binding"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]