On 02/10/2023 22:23, Balas, Eliza wrote: >>>> If you have version registers in these IPs, what benefit does version >>>> numbers in the compatible string bring? >>>> Rather than using the version numbers to validate what the DT gave you, >>>> which not the kernel's job IMO, why not just use the information from >>>> the register to determine the version? >>>> >>>> Cheers, >>>> Conor. >>> >>> As the description of this patch says, we want to resolve the naming confusion around the existing repurposed TDD core >> (https://wiki.analog.com/resources/eval/user-guides/ad-pzsdr2400tdd-eb/reference_hdl#tdd_controller) >>> built for AD9361 and this TDD Engine IP core (https://wiki.analog.com/resources/fpga/docs/axi_tdd) which is a similar core, with >> more output channels and some extra features. The version numbers in the compatible string are used to differentiate between the >> two IPs. >> >> Firstly, please fix your mail client to wrap text at a sane width :) >> Secondly, where is the binding for that TDD ad9361 specific core that >> calling this generic one "adi,axi-tdd" would conflict with? >> Grepping the bindings directory of the kernel tree for "adi.*tdd" returns >> nothing. If there is an ad9361 specific tdd, I would expect it to have a >> compatible like "adi,ad9361-tdd". > > We didn't upstream the ad9361 tdd driver, but we are using it > in our internal kernel. If this is an issue, I will change the > compatible string to "adi,axi-tdd". > Unfortunately, we do not care much about your internal kernel and whatever you have there should rarely be an argument in upstream discussions. Best regards, Krzysztof
