[PATCH v2 2/9] media: dt-bindings: add rockchip rk3588 vicap
Michael Riesch
michael.riesch at collabora.com
Thu Mar 19 01:04:06 PDT 2026
Hi Conor,
On 3/18/26 18:52, Conor Dooley wrote:
> [...]
>>
>> resets:
>> - items:
>> - - description: ARST
>> - - description: HRST
>> - - description: DRST
>> - - description: PRST
>> - - description: IRST
>> + minItems: 5
>> + maxItems: 9
>>
>> reset-names:
>> - items:
>> - - const: arst
>> - - const: hrst
>> - - const: drst
>> - - const: prst
>> - - const: irst
>> + minItems: 5
>> + maxItems: 9
>
>> +allOf:
>> + - if:
>> + properties:
>> + compatible:
>> + contains:
>> + const: rockchip,rk3568-vicap
>> + then:
>> + properties:
>> + clocks:
>> + minItems: 4
>> + maxItems: 4
>
> Anything here that repeats the outermost constraint should be dropped,
> so no minItems here..
I already took a look at similar VICAPs (e.g., R3576) that will have
more clocks. This is way I left this one in here.
But I guess I will remove it and add it again when e.g. RK3576 VICAP
support is added.
>> [...]
>
>> +
>> + reset-names:
>> + items:
>> + - const: arst
>> + - const: hrst
>> + - const: drst
>> + - const: irst0
>
> Were you willing to reuse "irst" for the 0th irst, all the repetition
> wouldn't be needed. I assume it represents the reset for the port at
> address 0, so it wouldn't even be a different meaning on each device.
Not quite I think. "{a,h,d}rst" are the same on all variants, then there
is "prst" and "irst" on RK3568 and "irst{0-5}" on RK3588. So I thought
that writing it out explicitly for each variant is more readable (even
though it may need a few characters more).
OK?
I'll try to send out a new version later today.
Best regards,
Michael
>
>> + - const: irst1
>> + - const: irst2
>> + - const: irst3
>> + - const: irst4
>> + - const: irst5
>> +
>> additionalProperties: false
>>
>> examples:
>>
>> --
>> 2.39.5
>>
>>
More information about the Linux-rockchip
mailing list