Skip to content
Snippets Groups Projects
Forked from postmarketOS / q6voiced
Up to date with the upstream repository.
user avatar
Arnaud Ferraris authored
Tinyalsa is an Android library, very rarely packaged by traditional
Linux distributions. It can easily be replaced with the more usual
`libasound` in this case, which will make packaging it easier.
c26193c6
History
Name Last commit Last update
LICENSE
README.md
meson.build
q6voiced.c

q6voiced

q6voiced is a userspace daemon for the QDSP6 voice call audio driver. Voice call audio is directly routed from the modem to the input/output audio devices, but something needs to start the audio streams for that to happen.

q6voiced listens on DBus system bus for signals from oFono and ModemManager, and opens/closes the PCM device when a DBus signal indicating that a call is initiated/ended is received. This essentially makes voice call audio work out of the box (provided that the audio routing, e.g. Earpiece and a microphone is set up appropriately).

The q6voiced patches can be currently found in the msm8916-mainline/linux repository (look for ASoC: qdsp6:) commits). It may also work for downstream since the currently implemented approach is very similar.

Note

It is expected that this daemon will be obsolete eventually. Different approaches exists for activating such audio streams that are not processed directly by Linux. (See Hostless PCM streams) At the moment the kernel driver implements the "Hostless FE" approach (as on downstream), but eventually this should be replaced by a Codec <-> Codec link.

In that case the audio streams would be activated by setting some ALSA mixers (e.g. through ALSA UCM) and this daemon could be removed. However, there is quite some work involved to make that work properly. Until then, this daemon allows voice call audio to work without activating audio manually.

Configuration

The q6voiced daemon expects a configuration file in /etc/conf.d/q6voiced. The file looks like this:

q6voice_card=$CARD
q6voice_device=$DEVICE

You can find the values for $CARD and $DEVICE using alsactl info from the alsa-utils package. You are looking for the modem audio device (example: VoiceMMode1) which is defined in your dts:

dts

voicemmode1-dai-link {
    link-name = "VoiceMMode1";
    cpu {
        sound-dai = <&q6voicedai VOICEMMODE1>;
    };
};

alsactl info output

- card: 0
  id: SHIFT6mq
  name: SHIFT6mq
  longname: SHIFT6mq
  mixer_name: 
  components: 
  controls_count: 1268
  pcm:
    - stream: PLAYBACK
      devices:
        ...
        - device: 12
          id: VoiceMMode1 (*)
          name: 
          subdevices:
            - subdevice: 0
              name: subdevice #0

In this case, the values are $CARD=0 and $DEVICE=12.