On 16.03.2023 23:47, Rob Herring wrote:
> On Wed, Mar 08, 2023 at 02:28:53AM +0100, Konrad Dybcio wrote:
>> The driver itself does not read the -cells values (and frankly, it
>> shouldn't),
>
> Agreed, because this is standard address translation and only the DT
> core address functions should read cells props.
>
>> so there's little sense in only allowing [1, 2] x [1].
>
> Why does the timer need 64-bits of address space? It doesn't, so that's
> the reason for restricting it.
>
>> Allow any values.
>>
>> Fixes: 4d2bb3e65035 ("dt-bindings: timer: Convert ARM timer bindings to json-schema")
>> Signed-off-by: Konrad Dybcio <konrad.dybcio@xxxxxxxxxx>
>> ---
>> .../devicetree/bindings/timer/arm,arch_timer_mmio.yaml | 6 ++----
>> 1 file changed, 2 insertions(+), 4 deletions(-)
>>
>> diff --git a/Documentation/devicetree/bindings/timer/arm,arch_timer_mmio.yaml b/Documentation/devicetree/bindings/timer/arm,arch_timer_mmio.yaml
>> index f6efa48c4256..236e2a05c1ad 100644
>> --- a/Documentation/devicetree/bindings/timer/arm,arch_timer_mmio.yaml
>> +++ b/Documentation/devicetree/bindings/timer/arm,arch_timer_mmio.yaml
>> @@ -26,11 +26,9 @@ properties:
>> maxItems: 1
>> description: The control frame base address
>>
>> - '#address-cells':
>> - enum: [1, 2]
>> + '#address-cells': true
>
> So 3 address cells is valid?
>
> Until recently (and not yet in a dtschema release), there was no
> constraint on #address-cells or #size-cells values other than the
> #.*-cells constraint of 8. Now it is 3 and 2.
Okay, I didn't know that. I suppose we can forget about this patch.
Konrad
>
>>
>> - '#size-cells':
>> - const: 1
>> + '#size-cells': true
>>
>> ranges: true
>>
>> --
>> 2.39.2
>>
