Hello all,
since kernel 3.7 I get hard locks on random CPUs when I plug in the charger.
Using openSUSE 12.2 on a Toshiba R840 with 3.7.1-1.1 kernel-desktop from Kernel stable repo atm.
And no it is not an hardware issue. It didn’t lock in BIOS and it didn’t lock with kernel <3.7.
After a second or two it switches to console with a kernel panic.
Any suggestions? More info needed?
m0nk
Are you by chance loading the VirtualBox drivers? I had a problem with the nouveau driver in kernel 3.7.x and somehow, updating the VirtualBox driver to 4.2.6 fixed the problem. I also see some complaints about kernel 3.7.1 being loaded from the Tumbleweed repository not working with the nVIDIA proprietary video driver.
Thank You,
Thank you for the suggestion, but neither VirtualBox nor Nvidia get loaded, Device has Intel graphics.
Yesterday I got a softlock during compilation of handbrake. I guess, I insert another HDD and install Windows to compare.
m0nk
Ok, I’ve forced a panic and took a photo.
https://www.bitcrazy.de/~ds/kernel_panic.jpg
Because it only happens on state change of battery/charger, it seems to be related to the timer/frequency/ticks, although I’m not a kernel developer.
m0nk
What file systems are you using? Is btrfs in there by chance?
Thank You,
On 12/25/2012 09:06 AM, jdmcdaniel3 wrote:
>
> dschr;2513205 Wrote:
>> Ok, I’ve forced a panic and took a photo.
>>
>> https://www.bitcrazy.de/~ds/kernel_panic.jpg
>>
>> Because it only happens on state change of battery/charger, it seems to
>> be related to the timer/frequency/ticks, although I’m not a kernel
>> developer.
>>
>> m0nk
>
> What file systems are you using? Is btrfs in there by chance?
It is probably not related to any file system. The photo is hard to read, but I
did see enough to know that the panic is caused by an attempt to kill the idle
task. How it got to that condition is well beyond my kernel debugging skills.
Post the conditions needed to cause the panic and a link to your photo on the
linux-kernel@vger.kernel.org mailing list. I would use “kernel panic” rather
than “hardlock” in the subject, as the latter means that the kernel is still
running, but has reached some sort of deadlock. A panic means that the kernel
has aborted.
Idle thread is likely red herring: ‘Bisected oops regression between v3.7-rc8 and v3.7: nmi_watchdog’ thread - MARC](http://marc.info/?t=135547776100001&r=1&w=2)
Yeah, exactly that’s it. Tried with
echo 0 > /proc/sys/kernel/nmi_watchdog
echo 1 > /proc/sys/kernel/nmi_watchdog
and it hung immediately.
Now trying to edit laptop-mode.
m0nk
