Hi,
Am 22.05.2019 um 19:28 schrieb Michael Mrozek EvilDragon@openpandora.org:
Well, audio is MOSTLY working, but needs a bit more work.
Maybe someone here does know where that stuff is defined and where you can fix it.
THE CURRENT STATUS:
We've done these tests with hns' image with Letux Kernel 4.19.
What is the issue: ALSA reports 7 output devices. Accessing some of them leads to a kernel lock.
Some are additional sound cards for: * bluetooth audio * LTE modem voice channel
These do not work if the respective clock is missing. This is not yet debugged and therefore never did work.
Maybe the device tree definitions are not (and never were) exactly correct (clock polarity or source for example or chip select). There may also be issues in the mcbsp driver.
I have no running Pyra with me so I can't check, but I think "aplay -l" reports the device names.
One of the outputs is the headset port - and that one plays audio smoothly without any issues, crackles, anything. It properly works!
However, speaker output does not work.
Speaker ootput is the same alsa device as the headset. But the other two of 4 channels.
The /root/twl script does some magic remixing the audio file contents to 4 channels.
Due to the fact that the speakers don't work, we have a lot of weird output devices and the headset works fine, our guess is that the devices are not properly setup in the TWL-Driver / boardfiles / whereever.
Does anyone know where this is all setup?
The key problem is the twl6040 driver and aess extensions. And the mcbsp/pdm stuff.
BR, Nikolaus