linux Unifi topology name fixup for Intel mach

This pull request tries to reuse the code which fixup I2S topology name to fixup HDA topology name as well for DMIC number.

Jun 04 '24 02:06 brentlu

There is no bug/issue here. I just try to use same code to fixup topology name of I2S/HDA machine drivers. And yes, we could use same code for SDW as well. Let me upload an new version after local testing. Thanks.

Jun 04 '24 07:06 brentlu

I've uploaded new patches which are focusing on the topology name fixup and minimizing each patch. Now all mach could use same code to handle DMIC name fixup (-dmicXch or -Xch). The flow to enumerate mach is not touched. Only the code related to topology name fixup is modified.

Jun 05 '24 11:06 brentlu

Not sure what's going on but we've got a nasty TGL IPC timeout I don't recall having seen before:

https://sof-ci.01.org/linuxpr/PR5037/build3377/devicetest/index.html?model=TGLU_UP_HDA-ipc4&testcase=check-signal-stop-start-capture-10

Jun 06 '24 16:06 plbossart

Not sure what's going on but we've got a nasty TGL IPC timeout I don't recall having seen before:

https://sof-ci.01.org/linuxpr/PR5037/build3377/devicetest/index.html?model=TGLU_UP_HDA-ipc4&testcase=check-signal-stop-start-capture-10

This PR only touch topology name string. Should not result in runtime IPC error. Maybe we could rerun the test?

Jun 11 '24 05:06 brentlu

SOFCI TEST

Jun 13 '24 14:06 plbossart

SOFCI TEST

Jun 17 '24 11:06 plbossart

Hi @plbossart @bardliao , is this PR looking good to you?

Jul 05 '24 00:07 brentlu

if the goal was to unify, why do we still need differences with SoundWire handling for the dmic case? it's the same in the end, we need the -2ch or -4ch suffix appended, no?

There is no functional change in these commits.

Are you sure? The firmware just stopped loading in some configurations with 2 DMICs:

#5124

Jul 30 '24 04:07 marc-hb