On 22/07/2022 08:02, Alexander Stein wrote: > Hello Krzysztof, > > thanks for your feedback. > > Am Donnerstag, 21. Juli 2022, 15:43:13 CEST schrieb Krzysztof Kozlowski: >> On 21/07/2022 15:33, Alexander Stein wrote: >>> Convert the NXP PCF85063 RTC binding to DT schema format. >>> >>> Add 'interrupts' and 'wakeup-source' as this device has an interrupt >>> which was not documented, but is in use. >>> 'clock-output-names' and '#clock-cells' are added as well, those were >>> probably missed when adding clkout support in commit 8c229ab6048b >>> ("rtc: pcf85063: Add pcf85063 clkout control to common clock framework") >> >> Thanks for adding it here, this sounds fine but brought my attention to >> interrupts and quartz-load. It seems that only rv8263 supports >> interrupts. In the same time rv8263 work only with 7000 >> quartz-load-femtofarads. >> >> If that's correct, you need to put "allOf" after "required" and inside >> "if:then:" restricting it. For rv8263 interrupts:true and quartz as >> const 7000, for else: interrupts:false. > > It is slightly different. In all the datasheets I found there was an IRQ pin, > so this applies to all models, although only some of them (PCF85063A, > PCF85073A and RV8263) support alarms, which is what Linux cares for right now. > But this is handles in the driver already. OK, this is fine then. > quartz-load-femtofarads does not apply to RV8263, because it has no OSCI pins > at all but uses an onboard oscillator. See commit 5b3a3ade0293 ("rtc: > pcf85063: add Micro Crystal RV8263 support") for that. But this also handled > in the driver already. This is what I was based on, so the quartz-load-femtofarads should not be even allowed for RV8263. > Apart from that apparently only PCF85063 has a fixed quartz-load of 7pF, the > other types supported can have either 7 oder 12.5 pF. ...and for PCF85063 this should be fixed to 7. Best regards, Krzysztof