On 25/03/2022 13:13, Roger Quadros wrote: > > > On 25/03/2022 14:08, Krzysztof Kozlowski wrote: >> On 25/03/2022 13:05, Roger Quadros wrote: >>> >>> >>> On 24/03/2022 20:26, Krzysztof Kozlowski wrote: >>>> On 23/03/2022 12:18, Roger Quadros wrote: >>>>> TI's AM64 SoC has the Error Locator Module. Add compatible and >>>>> related properties to support ELM on AM64 SoC. >>>>> >>>>> Signed-off-by: Roger Quadros <rogerq@xxxxxxxxxx> >>>>> --- >>>>> .../devicetree/bindings/mtd/ti,elm.yaml | 27 ++++++++++++++++++- >>>>> 1 file changed, 26 insertions(+), 1 deletion(-) >>>>> >>>>> diff --git a/Documentation/devicetree/bindings/mtd/ti,elm.yaml b/Documentation/devicetree/bindings/mtd/ti,elm.yaml >>>>> index 24ed0fb84204..bc01d35ce605 100644 >>>>> --- a/Documentation/devicetree/bindings/mtd/ti,elm.yaml >>>>> +++ b/Documentation/devicetree/bindings/mtd/ti,elm.yaml >>>>> @@ -17,7 +17,9 @@ description: >>>>> properties: >>>>> compatible: >>>>> items: >>>>> - - const: ti,am3352-elm >>>>> + - enum: >>>>> + - ti,am3352-elm >>>>> + - ti,am64-elm >>>>> >>>>> reg: >>>>> maxItems: 1 >>>>> @@ -25,6 +27,17 @@ properties: >>>>> interrupts: >>>>> maxItems: 1 >>>>> >>>>> + clocks: >>>>> + maxItems: 1 >>>>> + description: Functional clock. >>>>> + >>>>> + clock-names: >>>>> + items: >>>>> + - const: fck >>>>> + >>>>> + power-domains: >>>>> + maxItems: 1 >>>>> + >>>>> ti,hwmods: >>>>> description: >>>>> Name of the HWMOD associated with ELM. This is for legacy >>>>> @@ -37,6 +50,18 @@ required: >>>>> - reg >>>>> - interrupts >>>>> >>>>> +allOf: >>>>> + - if: >>>>> + properties: >>>>> + compatible: >>>>> + contains: >>>>> + const: ti,am64-elm >>>>> + then: >>>>> + required: >>>>> + - clocks >>>>> + - clock-names >>>>> + - power-domains >>>> >>>> Are these new properties also valid for am3352? >>> >>> No they are not required for am3352-elm. Only required for K3 based platforms like AM64. >>> >> >> I understand they are not required, but I asked if they are valid. IOW, >> whether ELM in am3352 also is part of power domain and has clock input? > > Yes it does have power and clock domains but they are handled a bit differently > on AM335x platform. i.e. in the parent node Then I guess it can stay like this. Otherwise (if there is no clock input) "else:" with "clocks:false" would be expected. Best regards, Krzysztof
