On 27/04/2025 05:55, Manikandan Karunakaran Pillai wrote: > work. >>>>>>> >>>>>>> Same applies to the other binding patch. >>>>>> Additionally, since this IP is likely in use on your sky1 SoC, why is a >>>>>> soc-specific compatible for your integration not needed? >>>>>> >>>>> The sky1 SoC support patches will be developed and submitted by the >> Sky1 >>>>> team separately. >>>> Why? Cixtech sent this patchset, they should send it with their user. >>> >>> Hi Conor, >>> >>> Please look at the communication history of this website. >>> >>> https://urldefense.com/v3/__https://patchwork.kernel.org/project/linux- >> pci/patch/CH2PPF4D26F8E1C1CBD2A866C59AA55CD7AA2A12@CH2PPF4D26F >> 8E1C.namprd07.prod.outlook.com/__;!!EHscmS1ygiU1lA!Gh- >> UeyTbbr2R3ocWWa4QZHM_GYBRXws7a5zc3lZvSy_XYVCkcg8mmeEaAWS4wEvI >> SMV2tGCEylE$ >> >> And in that thread I asked for Soc specific compatible. More than once. >> Conor asks again. >> >> I don't understand your answers at all. > > The current support is for the IP from Cadence. There can be multiple SoC developed based on this IP and it is for > the SoC companies to build in support as and when the SoC support needs to be available. > > Since the CIX SoC is available, it can be send together with this patch. > However, I do not understand the need for clubbing these in a single patch. No one asks for this. The point is such IP blocks are usually customized per SoC this generic compatibles are not enough. That's the argument here, not whether you can have multiple vendors (we all know this, imagine we know Cadence, Synopsys etc) or whether you want to combine here Cix or not. Answer rather how much software interface is compatible or common between different implementations. ... AND even then you always need soc specific compatible. See writing bindings for the reason (or any other tutorial/guide/speech about writing bindings). Best regards, Krzysztof
