Final update (hopefully): It seems that I have been able to fix the issue. I’m not sure what exactly caused the problem but either removing fluidsynth or installing the wireplumber ppa fixed the issue and I have working audio again. I’ve also removed pulseaudio and undid my edit to the modprobe blacklist, as they were only done as a temporary solution and they are no longer necessary.

For the past three days, I’ve been having this issue where my computer starts with no audio and the only sound device listed is a “dummy output” device. I’ve tried looking online for solution but the only solution I found has to be redone manually every time I start/restart my computer. It also seems like this issue is common with and possibly specific to the sound card my computer has, which is an “Intel Sunrise Point-LP HD Audio”.

The solution that worked for me was to add blacklist snd_soc_avs to the modprobe blacklist and then run the two commands sudo alsa force-reload and pulseaudio. Adding snd_soc_avs to the blacklist permanently brought back my actually audio devices but it didn’t fix the audio nor did it remove the dummy output device. The two commands I listed do restore the audio and remove the dummy output device but they only work for the current session and I have to run them again after starting/restarting my computer.

I have no problem doing this if there isn’t a permanent solution but I would like a permanent solution, if possible.

  • vortexal@lemmy.mlOP
    link
    fedilink
    arrow-up
    4
    arrow-down
    1
    ·
    4 days ago

    I thought that it was a kernel update too but the last time it was updated was two weeks ago and nothing else relevant to audio was updated either.

    I’ve restarted my computer and the fuser command does show mutiple instances of fluidsynth. I also ran the journalctl command and I’m getting a bunch of the same error messages over and over again. They are:

    mod.jackdbus-detect: Failed to receive jackdbus reply: org.freedesktop.DBus.Error.ServiceUnknown: The name org.jackaudio.service was not provided by any .service files

    spa.alsa: 'front:0': playback open failed: Device or resource busy

    mod.adapter: 0x5794b3a2b2a0: can't get format: Device or resource busy

    • gnuhaut@lemmy.ml
      link
      fedilink
      arrow-up
      3
      ·
      4 days ago

      Yeah actually looks like the same issue I had. Just uninstall fluidsynth with apt remove fluidsynth and (probably) reboot, and see if that fixes it.

      • vortexal@lemmy.mlOP
        link
        fedilink
        arrow-up
        1
        ·
        4 days ago

        Nope, that didn’t work. In fact, it made the issue worse because now I can’t get audio to work at all because my normal audio devices are missing again. I also tried running the commands again and the journalctl command is still giving me the same error messages and fuser states that the only thing running is pipewire.

        Also, for some reason, my computer took a longer time to boot than normal and it made me input my password at startup, which I have Linux Mint configured to just automatically log me in without it. So if you have any further suggestions that require a restart, I don’t feel comfortable restarting my computer again and I will try them tomorrow.

        • gnuhaut@lemmy.ml
          link
          fedilink
          arrow-up
          2
          ·
          edit-2
          4 days ago

          Oh, I’m sorry. Can you run systemd-analyze blame (and also systemd-analyze --user blame) and look for something, near the top, that took a long time? Also systemctl list-units --failed to see if something has failed to load properly.

          I’m afraid to ask, and I should have thought of that, do you remember if uninstalling fluidsynth removed anything else?

          Also maybe try uninstalling pulseaudio, it’s possible it’s fighting with pipewire over the device, but that’s just a guess.

            • gnuhaut@lemmy.ml
              link
              fedilink
              arrow-up
              2
              ·
              edit-2
              4 days ago

              Oh good. Maybe you can try restarting pipewire w/o rebooting.

              First, check if pulse is running (it might be even if uninstalled), with systemctl --user list-units (search for pulse by typing /pulse) or ps -A|grep pulse and kill it with (probably) something like systemctl --user stop pulseaudio or maybe killall pulseaudio. Not sure what’s better here, but it needs to be killed if it’s running.

              Then do systemctl --user restart pipewire pipewire-pulse wireplumber.

              • vortexal@lemmy.mlOP
                link
                fedilink
                arrow-up
                1
                ·
                4 days ago

                For some reason the the “stop” command didn’t work as it thought pulseaudio wasn’t running but I was able terminate it through htop. Also, that didn’t work, it did restart pipewire but my normal audio devices are still missing and I still don’t have working audio. I did, just in case, also check the journalctl and fuser commands you gave me previously, and fuser still just lists pipewire and journalctl stil gives the same error messages as before.

                • gnuhaut@lemmy.ml
                  link
                  fedilink
                  arrow-up
                  2
                  ·
                  4 days ago

                  Yeah I wrote that stop command wrong, it’s supposed to be systemctl --user stop pulseaudio. The relevant errors from journalctl were the “busy” errors. Are they still there? Maybe they’re old messages? You can type G to go to the end of the log.

                  Also, and this is probably my last suggestion, try restarting the socket with systemctl --user restart pipewire-pulse.socket. And maybe restart all the other crap as well for good measure. My theory here is that pulseaudio overwrote that socket with it’s own socket, and anything trying to play sound would be trying to connect to the nonexistent pulseaudio, and maybe restarting pipewire doesn’t actually recreate the socket, because systemd does that, but I’m not sure that’s actually how that works.

                  Theoretically logging out and in again should also restart all the things pipewire I think, but it’s possible that whatever is slowing down your boot is actually slowing down the login, so do at your own risk.

                  • vortexal@lemmy.mlOP
                    link
                    fedilink
                    arrow-up
                    1
                    ·
                    4 days ago

                    Ok, so a lot of them are old messages, none of the messages from this session are labeled as busy. I did just try logging out and back in and that was pretty much instantaneous, so whatever it was that caused my computer to boot slowly just effect the boot itself. But yeah, I tried restarting pipewire and everything related to it and it’s still just showing the dummy output device and audio isn’t working. Thanks for trying though.

          • vortexal@lemmy.mlOP
            link
            fedilink
            arrow-up
            1
            ·
            4 days ago

            I’m not sure how long everything takes normally but systemd-udev-settle.service took over two minutes. When running it with --user the longest was xdg-desktop-portal.service, which took 6 seconds. The third command gives:

            UNIT LOAD ACTIVE SUB DESCRIPTION ● casper-md5check.service loaded failed failed casper-md5check Verify Live ISO checksums ● systemd-udev-settle.service loaded failed failed Wait for udev To Complete Device Initialization ● vboxdrv.service loaded failed failed VirtualBox Linux kernel module Also, no I did check and fluidsynth was the only thing removed. I think it did for some reason add some of Wine’s dependencies to autoremove but I’ll deal with those later.

            • gnuhaut@lemmy.ml
              link
              fedilink
              arrow-up
              2
              ·
              4 days ago

              Yeah not sure about that. My guess is the casper-md5check is irrelevant, not sure why udev would be affected by installing/uninstalling audio stuff, and also not sure if that actually slows down your boot, my guess would be no. vboxdrv and virtualbox are probably also irrelevant.

              I think I should have included systemctl --user list-units --failed previously.