Thread (20 messages) 20 messages, 2 authors, 2025-09-12

RE: [PATCH v3 2/9] dt-bindings: media: nxp: Add Wave6 video codec device

From: Nas Chung <nas.chung@chipsnmedia.com>
Date: 2025-09-02 05:45:35
Also in: linux-devicetree, linux-media, lkml

Hi, Krzysztof.
-----Original Message-----
From: Krzysztof Kozlowski <krzk@kernel.org>
Sent: Friday, August 29, 2025 10:57 PM
To: Nas Chung <nas.chung@chipsnmedia.com>; mchehab@kernel.org;
hverkuil@xs4all.nl; robh@kernel.org; krzk+dt@kernel.org;
conor+dt@kernel.org; shawnguo@kernel.org; s.hauer@pengutronix.de
Cc: linux-media@vger.kernel.org; devicetree@vger.kernel.org; linux-
kernel@vger.kernel.org; linux-imx@nxp.com; linux-arm-
kernel@lists.infradead.org; jackson.lee [off-list ref];
lafley.kim [off-list ref]
Subject: Re: [PATCH v3 2/9] dt-bindings: media: nxp: Add Wave6 video codec
device

On 29/08/2025 10:46, Nas Chung wrote:
quoted
Add documents for the Wave6 video codec on NXP i.MX SoCs.
Pretty incomplete commit msg. Nothing explaining hardware, nothing
documenting resolution of previous discussions (where is all this
chip&media?).
I see,  I'll improve the commit message in v4 to include hardware details.
...

quoted
+
+properties:
+  compatible:
+    enum:
+      - nxp,imx95-vpu
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    maxItems: 1
+
+  power-domains:
+    maxItems: 1
+
+  memory-region:
+    maxItems: 1
+
+  sram:
+    $ref: /schemas/types.yaml#/definitions/phandle
+    description: phandle of the SRAM memory region node.
+
+  "#cooling-cells":
+    const: 2
+
+  "#address-cells":
+    const: 2
+
+  "#size-cells":
+    const: 2
+
+  ranges: true
+
+patternProperties:
+  "^video-core@[0-9a-f]+$":
+    type: object
Missing description.
I'll add a description in v4.
quoted
+    additionalProperties: false
+
+    properties:
+      compatible:
+        enum:
+          - nxp,imx95-vpu-core
Why do you need here compatible? Can this child be anything else? Can it
be re-used? Is it actually a separate block?

Your example suggests that the only distinctive resource are the
interrupt and address space and that's on the edge of calling it a
separate device.

There is some tendency to call such "pseudo-cores" a separate devices in
case of video codec bindings and experience shows these are usually
fake. It's not the same as DP or HDMI sub-block of display pipeline.

That's why you should come here with strong argument what separate piece
of hardware this is.
Thanks for your feedback.

As you mentioned, I wanted to represent the interrupts and address space
as separate "cores". This is because, from an external perspective (e.g. multi-VM),
each of these resources is a VPU interface and can be accessed independently
to operate the VPU.

However, there is indeed only one actual VPU processing engine.
I understand your point about "pseudo-cores".

I would appreciate any guidance on the preferred way to represent
these resources in the device tree.
quoted
+
+      reg:
+        maxItems: 1
+
+      clocks:
+        maxItems: 1
+
+      power-domains:
+        maxItems: 1
+
+      interrupts:
+        maxItems: 1
+
+    required:
+      - compatible
+      - reg
+      - clocks
+      - power-domains
+      - interrupts
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - power-domains
+  - memory-region
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+    #include <dt-bindings/clock/nxp,imx95-clock.h>
+
+    soc {
+      #address-cells = <2>;
+      #size-cells = <2>;
+
+      vpu: video-codec@4c4c0000 {
Unused label, drop
Okay. I'll drop the unused label.
quoted
+        compatible = "nxp,imx95-vpu";
+        reg = <0x0 0x4c4c0000 0x0 0x10000>;
+        clocks = <&vpu_blk_ctrl IMX95_CLK_VPUBLK_WAVE>;
+        power-domains = <&scmi_perf 10>;
+        memory-region = <&vpu_boot>;
+        sram = <&sram1>;
+        #cooling-cells = <2>;
+        #address-cells = <2>;
+        #size-cells = <2>;
+        ranges;
+
+        vpucore0: video-core@4c480000 {
None of these labels are used, drop.
I'll drop it.

Thanks,
Nas.
quoted
+          compatible = "nxp,imx95-vpu-core";
+          reg = <0x0 0x4c480000 0x0 0x10000>;
+          clocks = <&scmi_clk 115>;
+          power-domains = <&scmi_devpd 21>;
+          interrupts = <GIC_SPI 299 IRQ_TYPE_LEVEL_HIGH>;
+        };
+
+        vpucore1: video-core@4c490000 {
+          compatible = "nxp,imx95-vpu-core";
+          reg = <0x0 0x4c490000 0x0 0x10000>;
+          clocks = <&scmi_clk 115>;
+          power-domains = <&scmi_devpd 21>;
+          interrupts = <GIC_SPI 300 IRQ_TYPE_LEVEL_HIGH>;
+        };
+


Best regards,
Krzysztof
  
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help