[alsa-devel] [alsa-lib][PATCH] conf: SOF-Intel: map pcm device for sof driver

[Jaroslav Kysela]

Dne 16. 12. 19 v 13:51 Hui Wang napsal(a):
> On the machines with HDA codec, the sof driver will use the legacy hda
> codec driver, as a result, most mixer controls are same as before.
> 
> There are only 2 difference:
>   - HDMI pcm index is 3, 4 and 5 instead of 3, 7, 8...
>   - dmic is on device 6, 7 instead of 0
> 
> After applying this patch, we will get (aplay -L):
> hdmi:CARD=sofsklhdacard,DEV=0
>      sof-skl_hda_card,
>      HDMI Audio Output
> hdmi:CARD=sofsklhdacard,DEV=1
>      sof-skl_hda_card,
>      HDMI Audio Output
> hdmi:CARD=sofsklhdacard,DEV=2
>      sof-skl_hda_card,
>      HDMI Audio Output
> 
> and (arecord -L):
> dmic:CARD=sofsklhdacard,DEV=0
>      sof-skl_hda_card,
>      Digital Mic connected to the PCH directly
> dmic:CARD=sofsklhdacard,DEV=1
>      sof-skl_hda_card,
>      Digital Mic connected to the PCH directly
> 
> Then we could add minor changes in the pulseaudio configuration, after
> that, all output and input devices will work under pulseaudio.
> 
> This is a supplement for ucm, the ucm for sof driver has higher
> priority than this conf, if there is no ucm for sof driver, this conf
> will take effect.

Exactly. The dmic definitions should not be a standard in my eyes. I would not 
like to accept this. I can accept the standard definitions for the playback 
devices for this driver, but UCM will replace everything. It's better to 
define the special PA config for now until we resolve all the UCM issues like:

http://people.redhat.com/~jkysela/sof/carbon-x1-pulse.tar.gz2

The 2nd PA UCM patchset which is working with the current UCM in ALSA's repos 
for this hardware:

https://gitlab.freedesktop.org/pulseaudio/pulseaudio/merge_requests/217

But it is just a start. There are plenty unresolved abstraction issues with this.

						Jaroslav

-- 
Jaroslav Kysela <perex at perex.cz>
Linux Sound Maintainer; ALSA Project; Red Hat, Inc.