topology2: cavs-nocodec: Add an intermediate pipeline

[ranj063]

Add an intermediate pipeline between the host and DAI pipelines.

Signed-off-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com

[ranj063]

this is how the new tplg looks now.

[plbossart]

This is good to test the intermediate pipeline triggers but it should also be updated with the two-stage mixer if and when we figure out what to do with sink_c->sink and source_c->source

[aiChaoSONG]

does the mixout->gain->mixin pipeline have to be present only in playback? I think it is also possible in a capture pipeline.

[ranj063]

does the mixout->gain->mixin pipeline have to be present only in playback? I think it is also possible in a capture pipeline.

maybe but I guess for a test for intermediate pipelines, this is good enough

[kv2019i]

Almost 100% failure rate in Intel IPC4 plan, see test run 16208 . UPDATE: this can be unrelated as I see FW load failing, but needs to be confirmed.

[miRoox]

SOFCI TEST

[aiChaoSONG]

Almost 100% failure rate in Intel IPC4 plan, see test run 16208 . UPDATE: this can be unrelated as I see FW load failing, but needs to be confirmed.

I guess it is because this PR should be tested with the intermediate pipeline kernel support thesofproject/linux#3917

[lgirdwood]

Almost 100% failure rate in Intel IPC4 plan, see test run 16208 . UPDATE: this can be unrelated as I see FW load failing, but needs to be confirmed.

I guess it is because this PR should be tested with the intermediate pipeline kernel support thesofproject/linux#3917

@aiChaoSONG@kv2019i any update ? Can we merge ?

[lgirdwood]

@ranj063 looks like we have a conflict now.

[ranj063]

@ranj063 looks like we have a conflict now.

@lgirdwood this will have to wait until we bring back the mixer to the nocodec tplg. We're very close to doing that. Please hold off on merging this for a couple more days

[lgirdwood]

@ranj063 looks like we have a conflict now.

@lgirdwood this will have to wait until we bring back the mixer to the nocodec tplg. We're very close to doing that. Please hold off on merging this for a couple more days

ok, pls ping me when ready.

[lyakh]

how is this related to #6131 ? EDIT: ok, that one is topology1, sorry, ignore this

[plbossart]

the goal is the same @lyakh, for both topology1 and 2 we want to catch invalid firmware assumptions that a pipeline is connected to a dai or a host, and in the case of IPC4 detect cases where the intermediate pipeline is not correctly configured by the Linux driver.
Think of the nocodec topology as a 'pipe cleaner'

We have PR testing for IPC4, so as long as tests pass, I'm fine.

[kv2019i]

@ranj063 array-support merged in #7113 so rebase will be needed

[lyakh]

@ranj063 array-support merged in #7113 so rebase will be needed

@kv2019i wouldn't this be dropped in favour of #6981 or do we need both stream types in nocodec?