On 16/12/2024 11:54, Tomi Valkeinen wrote: > Hi, > > On 16/12/2024 12:42, Krzysztof Kozlowski wrote: >> On 16/12/2024 09:32, Laurent Pinchart wrote: >>> Hi Krzysztof, >>> >>> On Mon, Dec 16, 2024 at 08:58:49AM +0100, Krzysztof Kozlowski wrote: >>>> On Fri, Dec 13, 2024 at 04:02:59PM +0200, Tomi Valkeinen wrote: >>>>> From: Tomi Valkeinen <tomi.valkeinen+renesas@xxxxxxxxxxxxxxxx> >>>>> >>>>> The binding is missing maxItems for all renesas,cmms and renesas,vsps >>>>> properties. As the amount of cmms or vsps is always a fixed amount, set >>>>> the maxItems to match the minItems. >>>>> >>>>> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@xxxxxxxxxxxxxxxx> >>>>> --- >>>>> Documentation/devicetree/bindings/display/renesas,du.yaml | 10 ++++++++++ >>>>> 1 file changed, 10 insertions(+) >>>> >>>> The top level property should define widest constraints as well. >>> >>> I'm curious, why is that ? I understand why a top-level default would >>> make sense when it's optionally overridden by model-specific values, but >>> in this case there's no such default. Every SoC has its own fixed value. >> >> Because otherwise top level property does not have proper description >> and we expect properties to be defined at top-level. > > As we don't know what is the maximum number of items for future SoCs, > should we then use a number that'll surely be big enough? At the moment > the max cmms seems to be 4, so maybe 16 would be safely big enough. But > is it then better to be extra safe, and use, say, maxItems 256? No, look at all other bindings. Widest constraints for ONLY these devices. We do not talk about future SoCs here. Best regards, Krzysztof
