On Wed, Apr 29, 2020 at 10:24:45AM +0200, Arnd Bergmann wrote: > I looked at the bigger picture again and found that the more fundamental > problem is the dependency reversal in sound/soc/sof/sof-of-dev.c, where > you have common code that knows about and links against a hardware > specific driver. This is something we try hard do avoid in general in the > kernel, as it causes all kinds of problems: This is a legacy of this being factored out of the x86 code, since ACPI is not really fit for purpose when used to describe at least the audio hardware on modern laptops essentially all the enumeration is quirk based. It really needs cleaning up for the non-x86 SOF users.
Attachment:
signature.asc
Description: PGP signature