On Tue, Feb 20, 2018 at 11:15:01PM +0100, Hans de Goede wrote:
Instead of duplicating the DMI quirks between the codec and machine driver, add a rt5651_set_pdata() codec private function which the machine driver can use to pass in pdata.
It's not clear to me that this is a great idea. The goal in general is to keep as much of the CODEC configuration in the CODEC driver as possible so we can move more towards generic machine drivers, this is moving us in the opposite drection quite dramatically. That might get in the way of improvements we could be making when we get the component based changes further on and can avoid having all the DPCM stuff in the Intel drivers.
It is really unfortunate that ACPI based systems aren't embracing standards here and using this mess of open coded quirks but it feels like the way to do this is to keep the code as close as possible to well designed firmware interfaces and hope that they come round.