On 1/14/21 3:19 PM, Arnd Bergmann wrote:
On Thu, Jan 14, 2021 at 7:07 PM Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com wrote:
On 1/14/21 9:09 AM, Arnd Bergmann wrote:
On Tue, Jan 12, 2021 at 9:32 PM Arnd Bergmann arnd@kernel.org wrote:
v2: fix various build issues in the first version, now passes all randconfig builds I tried
Please disregard this version, while I have a tree that passes the randconfig builds now, this was not the patch from it that I wanted to send...
No worries, I reworked the PCI case completely. Still running tests to make sure there's no regression https://github.com/thesofproject/linux/pull/2683
Ok, I see. I had not realized that you already did the PCI bits as well, and applied my original patch to your tree, as I had meant to resend my "ASoC: SOF: ACPI: avoid reverse module dependency" patch after fixing some additional build failures in it. I have now applied the relevant commits from your branch to my randconfig tree, and will let you know if that finds anything more.
From what I can tell so far, you already included the fixups that I had locally, and more.
Thanks Arnd.
Yes we added a couple of things. I missed the fact that we hard-coded 'sof-acpi' to determine if we use the legacy or SOF driver, and other problems with driver names that I screwed-up.
The dependencies part should be finished now, I am just testing on my side that nothing broke on the test devices (which is slower than I wanted due to other ACPI scan things I had to revert locally). I will add a couple of suggestions from Guennadi but this should be done tomorrow.