topology: Add topology for Waves codec

[stolx]

Add pipe for Waves codec playback
Add pipe for Waves codec demux playback
Modify sof-tgl-max98357a-rt5682 topology so Waves codec can be added
to playback and headphones depending on -DWAVES value

Signed-off-by: Oleksandr Strelchenko oleksandr.strelchenko@waves.com

[stolx]

@cujomalainey @dbaluta @keyonjie
comments from #3951 are addressed, thank you

[cujomalainey]

Thanks for the update LGTM

[keyonjie]

looks good, thanks.

[plbossart]

couple of nit-picks below.

[plbossart]

does this mean that this will create controls "MaxxChrome Runtime %d" and "MaxChrome Setup %d"?

If yes, it's not clear to me what Runtime and Setup mean for a control? Adding some comments wouldn't hurt on the intended purpose and behavior

[cujomalainey]

These are controls dictated by the codec adapter and we have no control over.

[plbossart]

Sorry, what? Nothing prevents us from creating the controls we see fit in the topology. The codec adapter is just another module which knows nothing about controls. The codec adapter defines an IPC, but not how those controls are named and used.

[cujomalainey]

Agreed, but the names correspond to the naming in the codec adapter. Codec adapter has explicitly a "setup config" handler and a "runtime config" handler. I am down for comments I am just hesitant to deviate to far from the source code.

[stolx]

@plbossart

does this mean that this will create controls "MaxxChrome Runtime %d" and "MaxChrome Setup %d"?

yes, the topology will create:
for speakers: MaxxChrome Runtime 1 MaxxChrome Setup 1
for headphones: MaxxChrome Runtime 2 MaxxChrome Setup 2
difference between them is defined by codec_adapter api. For example, Setup, as opposed to Runtime, has a codec_adapter-related part that holds codec id (used to select between possible codecs), and a bunch of actually unused fields.
AFAIK having setup and runtime configs is a must for a codec that implements codec_adapter interface.
So correct description of Setup and Runtime is some kind of a reference to codec_adapter api, which I can add as a comment to topology but not sure it is needed.

Or maybe I did not understand your point correctly?

[plbossart]

the point was that it would be more intuitive to have

for speakers: MaxxChrome Runtime MaxxChrome Setup Speaker
for headphones: MaxxChrome Runtime MaxxChrome Setup Headphones

and as discussed with Curtis this can be done via macros that are set in the caller

[cujomalainey]

@stolx see here how all these params are defined in top level m4 then used as variables in included m4? @plbossart is suggesting something similar so we can pass in names to codec adapter where relevant. Just to make the names more intuitive.

[stolx]

@plbossart @cujomalainey
added macros to make more intuitive name for controls, please check if this will work

[marc-hb]

Github Actions hang fixed by #3992 , please git commit --amend and force push if you want to retrigger

[lgirdwood]

Github Actions hang fixed by #3992 , please git commit --amend and force push if you want to retrigger

@stolx can you make sure you rebase on top of latest main branch when pushing update as it include a CI fix (otherwise the CI shows many failures).

[plbossart]

looks mostly good, I suggested an improvement below and I have a question as well.
Thanks!

[plbossart]

I don't fully follow why here you have a PIPELINE ID but not in the other file without demux?

[stolx]

pipe-waves-codec-playback.m4 defines CA_RUNTIME_CONTROLBYTES_NAME and CA_SETUP_CONTROLBYTES_NAME
then sof/pipe-codec-adapter-playback.m4 adds pipeline ID to it:
https://github.com/thesofproject/sof/blob/main/tools/topology/sof/pipe-codec-adapter-playback.m4#L73

[stolx]

so here I am aligning with sof/pipe-codec-adapter-playback.m4 implementation adding pipeline ID to name of a control

[dbaluta]

@stolx this should be merged after #3999. Please mind that you need to define waves uuid
according to PR #3999

[lgirdwood]

@stolx #3999 is now merged.

[stolx]

@lgirdwood @dbaluta
I merged this PR with #3999 and added Waves codec UUID to custom topologies

[dbaluta]

@stolx thanks. That part looks good to me. For the rest I'm not able to comment as they are Intel specific stuff.

[lgirdwood]

SOFCI TEST

[lgirdwood]

Rerun CI as build reported a TGL-H failure, but status was all green ? Could be a CI sync issue. Rerun to be sure.

[lgirdwood]

Usual CI timeout on suspend/resume.