Comment 893 for bug 1958019

Revision history for this message
In , cam (cam-linux-kernel-bugs) wrote :

Created attachment 306213
kernel-mute.log

(In reply to Tintin from comment #874)
> # . tas2781-2dev-on.sh 2

Sorry if this is a bit long, but trying to include info to help debug the issue.

Well, it has finally happened again... It seems like the script hasn't really helped... But I'll have to try it again next time this happens to determine what, if any, role it had. So far it looks like it hasn't helped.

I was able to get my sound working by running this: "killall pipewire-pulse" and quitting/killing audacious (audacious usually but not always hangs when this starts happening), and quitting firefox, thought not necessarily in that order?

But even that isn't enough to completely fix it. Likely, it would come back as soon as I start playing audio from more than one application at a time.

When playing audio from another application, it would work for a few seconds, and then one of the applications would stop working. Almost like the audio device had become usable by one application at a time (like in the early ALSA days prior to pulseaudio becoming common). And probably if I stopped the other application, it usually wouldn't stop working again until repeating the above steps.

But playing with it some more (with both quitting applications, killing pipewire-pulse over and over), I was eventually able to get audio to consistently work again (ie, normally) for about the last 30 minutes or so.

I tried logging out and back at one point... That didn't seem to fix it either. It wasn't until several (or a bunch more) killing pipewire-pulse, restarting firefox and audacious a few times (I could try other software too!) that it started working again.

I'm leaning toward this being a pipewire-pulse issue at this point... but it's possible that there's a driver issue that causes pipewise-pulse becoming wedged, and that running tas2781-2dev-on.sh and killing pipewise-pulse I'm eventually able to get things into a good state.

If this happens again, I will conduct more testing... And I say if, because I think Ubuntu 24.04 is out today, and if it is... I will likely upgrade. If this is a software issue, good chance 24.04 won't have this problem. I'll report back with my results either way.

Anyone else tried killing pipewire-pulse when having issues?

I didn't see any sort of run away processes, I simply did "ps paux | grep pulse" and that was the only process that showed up so I tried killing it. It was just a guess from remembering all the early pulseaudio issues from over a decade ago. :D

Here's my info for convenience:
Kernel 6.8.7
Kubuntu 23.10
Running with KDE+Wayland

My pipewire-pulse info:

apt policy pipewire-pulse
pipewire-pulse:
  Installed: 0.3.79-2
  Candidate: 0.3.79-2
  Version table:
 *** 0.3.79-2 500
        500 http://us.archive.ubuntu.com/ubuntu mantic/main amd64 Packages
        100 /var/lib/dpkg/status

I didn't have this problem with the older kernel 6.7.x series, but maybe there was a pipewire-pulse update that happened? Or and update to a package it depends on?

Finally, I've attached the requested dmesg output. My audio problems started after 10:30 (maybe around 10:40-ish?) and my audio had been completely working for around 15 minutes as of the time I'd captured that dmesg output. Seems like nothing really relevant, but maybe there's something helpful?