[Bug 53216] Re: BUG() report in getnstimeofday: soft lockup

[gst]

It only solves the issue when chatting. For voice-calls it's still the
same problem.

So it seems to be pretty sure that this issue is related to audio output
(and maybe also input). But on the other hand the problem occurs with
the ALSA as well as the OSS kernel drivers.

Did anyone try if 2.6.20-vanilla solves this issue?

By the way - here is some more information on the replies which I
received to my LKML posting (http://lkml.org/lkml/2006/10/11/200) - I
just noticed that this mails weren't posted on the list. The discussion
is between another user experiencing the problem and one of the kernel
developers. Unlike this users system my system also completely hangs
with 2.6.19.

-------- --------- -------- --------
Currently I'm using 2.6.19-rc4 and Skype now behaves different. When calling,
the system load increases to 100% and audio stutters, but there is no lockup.
I'll test 2.6.19-rc5 during the weekend.
-------- --------- -------- --------
It'd be useful if you could grab a kernel profile when the system load is
high:

- Enable CONFIG_PROFILING

- Boot with `profile=1' on the kernel command line

- Run:

       readprofile -r
       sleep 10
       readprofile -n -v -m /boot/system.map | sort -n -k 3 | tail -40


Or, if oprofile is working:


#!/bin/sh
sudo opcontrol --stop
sudo opcontrol --shutdown
sudo rm -rf /var/lib/oprofile
sudo opcontrol --vmlinux=/boot/vmlinux-$(uname -r)
sudo opcontrol --start-daemon
sudo opcontrol --start
sleep 10
sudo opcontrol --stop
sudo opcontrol --shutdown
sudo opreport -l /boot/vmlinux-$(uname -r) | head -50
-------- --------- -------- --------
> Here is the oprofile log. Seems to be acpi related?
>
> CPU: CPU with timer interrupt, speed 0 MHz (estimated)
> Profiling through timer interrupt
> samples  %        symbol name
> 709      44.2848  acpi_pm_read
> 232      14.4909  schedule
> 164      10.2436  system_call
> 61        3.8101  __wake_up
> 48        2.9981  __copy_to_user_ll

I'd say skype is doing something dumb there.  The machine is apparently
madly context-switching and reading the time.
-------- --------- -------- --------
> I just upgraded to 2.6.19 and it's still the same behaviour (system
> hangs). Does it work for you when using 2.6.19 (non-rc)?

Yes, 2.6.19 works like the -rc kernels. I did not encounter any lockups
with these kernel. But I have to restart skype after some time to be
able to use audio. Otherwise, the audio quality is getting worse and
system load increases. I don't know what skype does, but maybe it is
related to skypes attempts to prevent debugging...
-------- --------- -------- --------

-- 
BUG() report in getnstimeofday: soft lockup
https://launchpad.net/bugs/53216