On Fri, Nov 10, 2017 at 02:34:31PM +0100, gregkh@xxxxxxxxxxxxxxxxxxx wrote: >On Thu, Nov 09, 2017 at 04:28:05PM +0000, Levin, Alexander (Sasha Levin) wrote: >> On Thu, Nov 09, 2017 at 04:17:33PM +0000, Mark Brown wrote: >> >On Thu, Nov 09, 2017 at 04:09:12PM +0000, Levin, Alexander (Sasha Levin) wrote: >> >> On Thu, Nov 09, 2017 at 11:07:44AM +0000, Mark Brown wrote: >> > >> >> >This doesn't seem like a good idea - note how the changelog says that >> >> >the transmit FIFO is at a different address, I'd expect this means that >> >> >people won't be able to use the new compatible without a corresponding >> >> >code change. >> > >> >> Interesting. I keep trying to automate DT patch selection to deal with >> >> issues as this, but it seems like there is always something... >> > >> >> I agree with Mark. I think it would make sense to pick 1bd92af877 as >> >> well to deal with this issue. >> > >> >Yeah, if they both go in together it's fine by me (though it's a little >> >more than just a device ID). >> >> It's a quirk too :) > >Ok, but 1bd92af877 does not apply cleanly (or even not cleanly) to 4.9, >can someone provide me a backport? Sure, please pick the following 3 commits, should be no conflicts then: - 96e53c41e1f81c9e9d1ce38d3f28b95668b71dcf - just dead code removal, avoids conflicts later. - 7762681a3ada5fca6017e75ea7f9cdac08fc50b9 - adds a basic quirks structure. - 1bd92af877abfeddcc4b83a35482ed4139591acf - adds support for H2 SoC. -- Thanks, Sasha