Re: Strange USB audio behavior

  Christos Margiolis wrote:

> On Fri Jun 19, 2026 at 7:55 PM EEST, Roman Bogorodskiy wrote:
> > Hi,
> >
> > For about a month I'm having a strange issue with the USB audio. I did
> > maybe 1 or 2 updates of -CURRENT during this time, but not sure when it
> > has started exactly as it does not show up immediately.
> >
> > The issue is: I have a set of USB speakers which I use as the main
> > output.
> >
> > After a few days of uptime (this time it was 3), the audio simply
> > disconnects.
> >
> > In dmesg it looks like:
> >
> > ugen0.4: <Generic Philips SPA 20/93> at usbus0 (disconnected)
> > umass0: at uhub0, port 9, addr 3 (disconnected)
> > da0 at umass-sim0 bus 0 scbus6 target 0 lun 0
> > da0: <Generic Masstorage 1.00>  s/n 20220701093627 detached
> > (da0:umass-sim0:0:0:0): Periph destroyed
> > umass0: detached
> > uaudio0: at uhub0, port 9, addr 3 (disconnected)
> > pcm2: detached
> > uaudio0: detached
> > usbhid1: at uhub0, port 9, addr 3 (disconnected)
> > hcons0: detached
> > hidbus1: detached
> > usbhid1: detached
> >
> > At this point, things appear quite strange. If I reconnect this device,
> > it's not detected anymore, complete silence. Moreover, if I connect
> > other USB devices, there are ignored by the system (no output in dmesg).
> >
> > The system has a builtin audio, but when this happens, switching to it
> > does not help:
> >
> >> cat /dev/sndstat
> > Installed devices:
> > pcm0: <Realtek ALC897 (Rear Analog)> (play/rec) default
> > pcm1: <Realtek ALC897 (Analog)> (play/rec)
> > No devices installed from userspace.
> >
> > But I get no audio when I try to test it with e.g. 'mpv -ao oss'.
> >
> > That could be a hardware issue, but feels more like a bug in the
> > USB/sound subsystem? Any ideas how to debug that?
> >
> > I'm running -CURRENT @ 0010c4b8a020dbeb81e61b71117d1caae9b044cc now.
> >
> > Roman
> 
> This could very well a cable or port issue. It's definitely not a sound
> driver problem because the whole USB device disconnects, not just the
> audio parts.
> 
> Christos

FWIW, one month later, I've updated -CURRENT twice and didn't have this
issue since than. I assume it was some kind of issue in -CURRENT which
is resolved now apparently.

It does not seem like a hardware issue because I think it would
keep happening at the roughly same rate as before. Will see if it shows
up again.

Roman