[RFC][HACK]: set alh channel = 2 by default

[bardliao]

To avoid it is reset to 0 when second alh is registered to the same
stream.

Signed-off-by: Bard Liao yung-chuan.liao@linux.intel.com

[lgirdwood]

@bardliao can you describe the problem here in more detail. i.e. include all the steps prior to the issue and how this works around the issue. @lbetlej fyi

[slawblauciak]

@lgirdwood the stream params are not properly propagated through the mux component. This hack should not be implemented, rather, the underlying problem should be fixed instead.
I'm currently investigating and looking for a proper way to fix this.

[lgirdwood]

@slawblauciak yep, thanks, that's what I was trying to establish here It seems we have some issues with MUX atm..

[bardliao]

@bardliao can you describe the problem here in more detail. i.e. include all the steps prior to the issue and how this works around the issue. @lbetlej fyi

The problem is that I will see comp_err(dev, "dai_params() error: invalid period_bytes."); when I try the synchronize playback topology. And the error is due to dd->local_buffer->stream.channels is 0.
Note that the two DAIs which connect to the demux share the same sof_ipc_stream_params{}. So, when alh_get_hw_params() is called for the 2nd DAI, the configured params will be reset to 0.

[lgirdwood]

@bardliao can you describe the problem here in more detail. i.e. include all the steps prior to the issue and how this works around the issue. @lbetlej fyi

The problem is that I will see comp_err(dev, "dai_params() error: invalid period_bytes."); when I try the synchronize playback topology. And the error is due to dd->local_buffer->stream.channels is 0.
Note that the two DAIs which connect to the demux share the same sof_ipc_stream_params{}. So, when alh_get_hw_params() is called for the 2nd DAI, the configured params will be reset to 0.

Why and where are they reset to 0 ? Can you describe the flow and topology.

[bardliao]

Why and where are they reset to 0 ? Can you describe the flow and topology.

The topology is #2542 and below is the flow

ipc_stream_pcm_params() 
    -> pipeline_params()
    -> loop calling pipeline_comp_hw_params() for each component 
    -> dai_hw_params() 
    -> alh_get_hw_params() where params->channels will be set to 0

Btw, I think @slawblauciak is looking into this issue.

[slawblauciak]

@lgirdwood when the values are set to 0, they are supposed to be set properly later, when stream parameters are propagated through the pipeline. However, that seems to be somewhat broken with mux/demux, I managed to pinpoint the issue. I'm currently working on a fix.

[bardliao]

@slawblauciak will create another PR to fix it.

[singalsu]

The rate should be set also (48000). Components SRC and ASRC need it.

This similar hack is also needed for HDA.