Let's see if this helps with
#6976
#6975

An alternative to revert of #6978

FYI @lgirdwood @mwasko @makarukp @mwasko @jsarha @marc-hb

CI is almost clean. There is one curious failure with alsabat on a TGL HDA device that I can't explain -> https://sof-ci.01.org/sofpr/PR6979/build3466/devicetest/index.html?model=TGLU_UP_HDA_IPC4ZPH&testcase=check-alsabat-headset-playback

The capture wave file is silent, but if one looks at the mixer settings during test, it seems the master-playback switch is off:

numid=2,iface=MIXER,name='Master Playback Switch'
  ; type=BOOLEAN,access=rw------,values=2
  : values=off,off

What I can't explain is why tests has been testing earlier today on the same device.

@makarukp @mwasko looks back to normal with this, we can work on the TGL integration as an incremental PR.

ADLP_RVP_NOCODEC_IPC4ZPH was "not available" in https://sof-ci.01.org/sofpr/PR6979/build3466/devicetest/index.html. Re-running it.

There's a WARNING: Signal too weak! failure in
https://sof-ci.01.org/sofpr/PR6979/build3466/devicetest/index.html?model=TGLU_UP_HDA_IPC4ZPH&testcase=check-alsabat-headset-playback

It maybe specific to sh-tglu-up-hda-01, @fredoh9 could you check?

SOFCI TEST

EDIT: ADLP_RVP_NOCODEC_IPC4ZPH still "not available" in https://sof-ci.01.org/sofpr/PR6979/build3474/devicetest/index.html

We had scheduled Jenkins maintenance that took longer than expected and there is backlog now, so maybe that's expected. Should SOFCI TEST again later.

EDIT: ADLP_RVP_NOCODEC_IPC4ZPH still "not available" in https://sof-ci.01.org/sofpr/PR6979/build3474/devicetest/index.html

I finally got to the bottom of this. Sort of. Both times, the test tried to run on sh-adlp-rvp-nocodec-06. But it was in a super weird state where it would successfully execute many ssh commands but fail to download anything or write anything to log files. Filed as internal issue 350. I rebooted it, let's see what happens now.

SOFCI TEST

https://sof-ci.01.org/sofpr/PR6979/build3475/devicetest/index.html has just one and unfortunately common suspend/resume failure, everything else is green.

(it used sh-adlp-rvp-nocodec-05 because sh-adlp-rvp-nocodec-06 was still reserved by me)