On 09/04/2022 14:55, Paul Cercueil wrote: >> >> I saw the efuse children and that's why I asked who is expected to >> populate them. You said that simple-mfd is required for this, I say >> no. >> It should work without simple-mfd... >> >> I am kind of repeating myself but I really do not see the need of >> simple-mfd in the bindings. > > Well, it is a "simple MFD", It's not a simple MFD, it is a memory controller. MFD is a purely Linux/software term, so there are no devices which are MFD. Everything which we model as MFD is actually something else in real life (e.g. PMIC, memory controller, system controller). > so I don't see why we can't use the > "simple-mfd" compatible. Why would we not want to use it? No one said that you cannot. You just might not need... > > Besides, if the nemc driver is responsible for populating the efuse > device, that means the nemc driver must be enabled for the efuse to > work, which is nonsense, the two IP blocks being unrelated. That's actually the explanation I was looking for. It would be nice to use it in commit msg instead of the dtbs_check warning. Best regards, Krzysztof
