On 20/05/2024 16:56, Eddie James wrote: > > On 5/19/24 13:03, Krzysztof Kozlowski wrote: >> On 14/05/2024 22:54, Eddie James wrote: >> >>> +properties: >>> + compatible: >>> + enum: >>> + - ibm,i2c-fsi >>> + >>> + reg: >>> + items: >>> + - description: FSI slave address >>> + >>> + "#address-cells": >>> + const: 1 >>> + >>> + "#size-cells": >>> + const: 0 >>> + >>> +patternProperties: >>> + "^i2c(@.*)?": >> Either you have or you have not unit addresses. Please fix the pattern. >> Why is this so flexible? Do you want to deprecate i2c-bus in favor of >> i2c? If so, then example should use new naming. I am fine with children >> as i2c-bus, assuming this is allowed by dtschema. Did you actually test it? > > > This is the exact pattern of the i2c-controller schema node name, which > I thought would be good. I can make it more specific. But yes I tested > it, i2c-bus works fine i2c-controller schema is a generic schema, applicable to various cases, including non-MMIO. You have here specific instances. Best regards, Krzysztof