topology2: add cpc value for gain

[aiChaoSONG]

The missing cpc in gain component results in wrong calculation for clock selection, and causes capture issues.

Link: #7228
Link: #7230

[abonislawski]

Thanks @aiChaoSONG can you also update other modules to match the manifest data?

[aiChaoSONG]

@RanderWang @abonislawski CPC for all modules in manifest are zero, but we also have CPS there, is it ok for you to calculate CPC from the maximum CPS for the module, to make CPC work for all supported formats.

I see several formula, could you please clarify which one is correct?

// from python tests
cps = cpc * sampling frequency // sample group in ms * frame size
cps = (cpc / (frame_size * frame_count)) * 1000

From the cps and cpc values for SRC moudle in mtl.toml, the formula look like
cps = cpc * 1000

The value 10183 for gain you write here #7228 (comment) seems not aligned with any of the above formula. how is it calculated?

[abonislawski]

@aiChaoSONG please use mtl-004 manifest as a reference:
https://github.com/thesofproject/rimage/commits/mtl-004-drop-stable
it contains updated manifest with measured values

@kfrydryx could you please clarify cps/cpc calculations?

[kfrydryx]

@abonislawski @aiChaoSONG CPS means cycles per second and CPC - cycles per chunk. And since we have 1ms scheduling period, CPS = CPC*(1s/scheduling_period) = CPC*1000. You can calculate the value in multiple ways, but the result will be the same.

[kv2019i]

FYI @mwasko @lgirdwood -- the relation to topology (and how to calculate) needs to be covered in sof-docs (related to thesofproject/sof-docs#451 )

[kv2019i]

https://sof-ci.01.org/sofpr/PR7242/build4478/devicetest/index.html hit a case of #7184 . Not related to this PR. Proceeding with merge.