On 30/03/2024 0:02, Krzysztof Kozlowski wrote: > > On 29/03/2024 14:36, Mark Brown wrote: > > On Fri, Mar 29, 2024 at 12:42:22PM +0100, Krzysztof Kozlowski wrote: > > > >> I stated and I keep my statement that such block is usually not > >> usable on its own and always needs some sort of quirks or > >> SoC-specific implementation. At least this is what I saw in other > >> similar cases, but not exactly I2S. > > > > I wouldn't be so pessimistic, especially not for I2S - a good portion > > of quirks there are extra features rather than things needed for basic > > operation, a lot of things that might in the past have been quirks for > > basic operation are these days hidden behind the DT bindings. > > OK, I trust your judgement, so cdns as fallback seems okay, but I don't think it > warrants cdns as used alone. Not particularly because cdns is different, but > because we expect specific SoC compatible always. > > Thus if cdns,i2s-mc stays, then: > > items: > - enum: > - starfive,jh8100-i2s > - cdns,i2s-mc > OK, thanks Krzysztof and Mark. I will modify it in next patch. Best regards, Xingyu Wu
