[PATCH 2/9] media: dt-bindings: add rockchip rk3588 vicap

[Michael Riesch]

Hi Conor,

On 3/13/26 17:57, Conor Dooley wrote:
> On Fri, Mar 13, 2026 at 04:56:29PM +0000, Conor Dooley wrote:
>> On Fri, Mar 13, 2026 at 04:20:44PM +0100, Michael Riesch via B4 Relay wrote:
>>> From: Michael Riesch <michael.riesch at collabora.com>
>>>
>>> Add documentation for the Rockchip RK3588 Video Capture (VICAP) unit.
>>>
>>> Signed-off-by: Michael Riesch <michael.riesch at collabora.com>
>>> ---
>>>  .../bindings/media/rockchip,rk3588-vicap.yaml      | 256 +++++++++++++++++++++
>>>  MAINTAINERS                                        |   1 +
>>>  2 files changed, 257 insertions(+)
>>>
>>> diff --git a/Documentation/devicetree/bindings/media/rockchip,rk3588-vicap.yaml b/Documentation/devicetree/bindings/media/rockchip,rk3588-vicap.yaml
>>> new file mode 100644
>>> index 000000000000..7fd4214921cb
>>> --- /dev/null
>>> +++ b/Documentation/devicetree/bindings/media/rockchip,rk3588-vicap.yaml
>>> @@ -0,0 +1,256 @@
>>> +# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
>>> +%YAML 1.2
>>> +---
>>> +$id: http://devicetree.org/schemas/media/rockchip,rk3588-vicap.yaml#
>>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>>> +
>>> +title: Rockchip RK3588 Video Capture (VICAP)
>>> +
>>> +maintainers:
>>> +  - Michael Riesch <michael.riesch at collabora.com>
>>> +
>>> +description:
>>> +  The Rockchip RK3588 Video Capture (VICAP) block features a digital video
>>> +  port (DVP, a parallel video interface) and six MIPI CSI-2 ports. It receives
>>> +  the data from camera sensors, video decoders, or other companion ICs and
>>> +  transfers it into system main memory by AXI bus and/or passes it to the image
>>> +  signal processing (ISP) blocks.
>>> +
>>> +properties:
>>> +  compatible:
>>> +    enum:
>>> +      - rockchip,rk3588-vicap
>>
>> Curious why this cannot share a binding with the existing 3568-vicap.
>> Looks pretty similar binding wise at least.
>> If it's an entirely different architecture or whatever, please mention
>> that in your commit message.
> 
> Looking further, it's using the same driver too...

It's not an entirely different architecture (indeed it uses the same
driver). There are some differences to the RK3568 and the PX30 (which
uses its own binding as well BTW): apart from different resets and
clocks that's mostly the notion of the connections to the ISP. But to be
fair, as it turns out this boils down to two additional ports.
Other recent SoCs (e.g., RK3576 or RK3562) will be a good match for the
RK3588 binding, but then again exactly resets, clocks and ports may vary
in that variants as well.

Personally I find this variant-specific DT binding magic hard to read,
and thus I went for a separate binding. That said, please let me know
what your preference is and I'll arrange it that way. Not a hill I'd
want to die on.

Best regards,
Michael