Bug#850806: [pulseaudio] pulseaudio hangs in "context_autospawn"

[Felipe Sateler]

Control: retitle -1 pulseaudio: hangs if dbus is malfunctioning
Control: tags -1 - moreinfo

On 10 January 2017 at 15:51, Bogdan Vatra <bogdan.vatra at kdab.com> wrote:
> On marți, 10 ianuarie 2017 20:01:29 EET Bogdan Vatra wrote:
>> On marți, 10 ianuarie 2017 19:47:44 EET Bogdan Vatra wrote:
>> > Hi,
>> >
>> > On marți, 10 ianuarie 2017 14:21:02 EET Felipe Sateler wrote:
>> > > On 10 January 2017 at 12:59, Bogdan Vatra <bogdan.vatra at kdab.com> wrote:
>> > > > Hi,
>> > > >
>> > > > Attached, but there is not to much in it ...
>> > > >
>> > > > BTW, immediately aster I'm installing pulseaudio a few (3-4) instances
>> > > > are
>> > > > spawn instantly.
>> > >
>> > > Wow, that hangs pretty fast. I'm guessing it hangs while trying to
>> > > acquire high-priority scheduling.
>> > >
>> > > The initial bug did not report the architecture, which one it is? Are
>> > > you using a debian-provided kernel or a custom-built one?
>> > >
>> > > Could you please (after installing pulseaudio-dbgsym), run under gdb
>> > > and present the trace of where it hangs? Note this is different from
>> > > the backtrace you already showed, as that is from the client, and we
>> > > want the one from the server.
>> >
>> > I strace it and judging from the log it doesn't seem to hang.
>> > Anyway I attached for you the gdb log as you requested.
>>
>> Sorry, I forgot to answer the arch & kernel questions :)
>> I'm using debian amd64 with debian-provided kernel (I'm way too lazy to
>> build my own).
>>
>> BogDan.
>
> Ah, I was wrong, it ~hangs (actually it waits 25s for every call) in
> rtkit_make_high_priority, when it calls
> dbus_connection_send_with_reply_and_block ....
> and it happens because rtkit fails to start :(

Right, so it appears the cause is rtkit failing. Still, pulseaudio
should not hang when rtkit cannot start.

As a workaround, you might try disabling high-priority and
realtime-scheduling in /etc/pulse/daemon.conf (although that will
probably give you suboptimal performance).

Could you please report this bug upstream at
https://bugs.freedesktop.org/enter_bug.cgi ?

-- 

Saludos,
Felipe Sateler