kmod test failed on some platforms with SST config enabled

[xiulipan]

After the SST config thesofproject/kconfig#42 merged
Most kmod test failed https://sof-ci.01.org/linuxpr/PR2511/build4725/devicetest/

rmmod: ERROR: Module snd_soc_acpi_intel_match is in use by: snd_soc_skl

[marc-hb]

Here are some (large) CI failures that seem related: basically any PR in any repo today. Feel free to edit my comment to add more.

https://sof-ci.01.org/softestpr/PR443/build307/devicetest
https://sof-ci.01.org/sofpr/PR3540/build7490/devicetest/
https://sof-ci.01.org/sofpr/PR3540/build7493/devicetest/
https://sof-ci.01.org/linuxpr/PR2518/build4731/devicetest/
https://sof-ci.01.org/softestpr/PR430/build305/devicetest/

[marc-hb]

https://sof-ci.01.org/sofpr/PR3540/build7493/devicetest/ says

21/10/2020, **14:10** (UTC)
Linux: f8922c96 ASoC: rt711-sdca: Add RT711 SDCA vendor-specific driver

But the kernel commit f8922c96 does not include the next Linux commit 37740626ab which was merged at 2020-10-21 02:21 UTC as shown by TZ=UTC git log --date=iso-local --format=fuller sof/topic/sof-dev and the merged time shown in thesofproject/linux#2487

So this SOF PR that ran at 14:10 UTC didn't include a 02:21 UTC kernel commit. I think this is proof that non-kernel Pull Requests use a nightly build of the kernel (for performance reasons?). @xiuli confirm?

If confirmed then this is enough to explain why most PRs failed today.

Warning: the SHA1s at https://sof-ci.01.org/ are incomplete because the PR's SHA1 is actually combined with the latest-branch-of-the-moment before testing (which is good) but there's no indication of what exactly it was combined with (which is bad). I filed internal issue 498 to address this.

[marc-hb]

Newer https://sof-ci.01.org/sofpr/PR3540/build7494/devicetest is mostly OK.