Finally cracked it!
Posted: Fri Feb 18, 2022 3:41 am
For a very long time the OpenFrame images haven't been able to move on from the 3.16 kernel branch due to an annoying issue with ALSA that has prevented audio hot-plugging from working. Well, to be fair, it's an annoying issue with how the STAC9202 chip is wired up in the Joggler / OpenFrame 1, but changes in ALSA were making all of the recommended fixes fail miserably.
It's been a pet problem of mine, particularly as one of the main uses of the Joggler has been as an audio playback unit. With the bug present, even when a cable from the back of the unit is permanently connected to headphones or an amplifier there would be an annoying crackling interference from the internal speakers.
WELL NO MORE!
I finally found a little time to work on Debian Bullseye (though I'd originally been targeting Buster... shows how much time has gone by) and in the process compiled the 5.10 kernel. As I checked the boot output for errors I noticed that it enumerated the STAC9202 IC in an identical manner to the 3.16 kernel. Weird, I thought, I'm sure it used to do that differently in the 4.x branch.
So I modified the kernel and userspace patches to apply to kernel 5.10 and tada! It behaves precisely as it did under 3.16. THE FINAL LEVEL COMPLETE!
Well, I'm not quite done, there's something weird happening with the backlight driver, but I think it's a simple fix.
With any luck there'll be some sort of Debian Bullseye image with an up-to-date and fully working kernel available in the next week or two.
It's been a pet problem of mine, particularly as one of the main uses of the Joggler has been as an audio playback unit. With the bug present, even when a cable from the back of the unit is permanently connected to headphones or an amplifier there would be an annoying crackling interference from the internal speakers.
WELL NO MORE!
I finally found a little time to work on Debian Bullseye (though I'd originally been targeting Buster... shows how much time has gone by) and in the process compiled the 5.10 kernel. As I checked the boot output for errors I noticed that it enumerated the STAC9202 IC in an identical manner to the 3.16 kernel. Weird, I thought, I'm sure it used to do that differently in the 4.x branch.
So I modified the kernel and userspace patches to apply to kernel 5.10 and tada! It behaves precisely as it did under 3.16. THE FINAL LEVEL COMPLETE!

Well, I'm not quite done, there's something weird happening with the backlight driver, but I think it's a simple fix.
With any luck there'll be some sort of Debian Bullseye image with an up-to-date and fully working kernel available in the next week or two.
