[alsa-devel] snd_usb_usx2y holding onto a lock?
Joseph Spencer
joespenceris at gmail.com
Sun Aug 19 22:27:19 CEST 2018
Thanks for the reply! I've been pouring over the source. Looks like it
really hasn't been actively maintained since around 2005, mostly minor
refactoring since then.
If I have questions that the docs available online can't answer, would this
be a good place to ask?
I appreciate all the work you guys do!
Joe Spencer
602.777.2307
On Sun, Aug 19, 2018 at 1:08 AM Takashi Iwai <tiwai at suse.de> wrote:
> On Fri, 17 Aug 2018 09:23:35 +0200,
> Joseph Spencer wrote:
> >
> > When plugging an audio interface in with nomodeset I see the following in
> > the logs:
> >
> > Aug 08 23:19:23 localhost.localdomain kernel: INFO: task pulseaudio:1750
> > blocked for more than 120 seconds.
> > Aug 08 23:19:23 localhost.localdomain kernel: Tainted: G
> > OE 4.17.11-200.fc28.x86_64+debug #1
> > Aug 08 23:19:23 localhost.localdomain kernel: "echo 0 >
> > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> > Aug 08 23:19:23 localhost.localdomain kernel: pulseaudio D12408
> 1750
> > 1268 0x00000000
> > Aug 08 23:19:23 localhost.localdomain kernel: Call Trace:
> >
> > Aug 08 23:19:23 localhost.localdomain kernel: ? __schedule+0x2ba/0xaf0
> >
> > Aug 08 23:19:23 localhost.localdomain kernel: ?
> > prepare_to_wait_event+0xd2/0x180
> > Aug 08 23:19:23 localhost.localdomain kernel: schedule+0x2f/0x90
> >
> > Aug 08 23:19:23 localhost.localdomain kernel:
> usX2Y_urbs_start+0x208/0x470
> > [snd_usb_usx2y]
> > Aug 08 23:19:23 localhost.localdomain kernel: ? finish_wait+0x80/0x80
> (snip)
>
> So this is waiting for the event to be triggered by the urb
> completion, but it didn't happen by some reason on your machine.
>
> > The audio interface doesn't work and without nomodeset X freezes.
> >
> > Any help pointing me in the right direction would be greatly appreciated.
> > It looks like an issue with the snd_usb_usx2y driver to me, but I'm not
> > sure.
>
> Unfortunately the driver is very special and does somewhat magical way
> only the author knows of well.
> You need to follow the code, and at least check whether the urb
> submission at the playback start worked properly as expected.
>
>
> Takashi
>
More information about the Alsa-devel
mailing list