2.6.30 kernel panic when closing lid

2.6.30 kernel panic when closing lid

[Andreas Sundstrom]

Hi..

My laptop has started crashing when I close the lid. This does not
happen with 2.6.29.4 which I was using before 2.6.30, I configured my
2.6.30 kernel with the 2.6.29.4 as the base and then did "make
oldconfig" and then most of the questions I got I used the default setting.

I have tried to catch the full output with a USB serial console, but
it's not able to send the full output before it dies it seems. Sometimes
it gets further than other times.

Here is one oops/panic:
[   31.295898] kernel tried to execute NX-protected page - exploit
attempt? (uid: 0)
[   31.296575] BUG: unable to handle kernel paging request at
ffff88013a891630
[   31.297233] IP: [<ffffffff814f6940>] fair_sched_class+0x0/0xc0
[   31.297594] PGD 202063 PUD b067 PMD 800000013a8001e3
[   31.298567] Oops: 0011 [#1] SMP
[   31.298569] last sysfs file:
/sys/devices/pci0000:00/0000:00:1c.1/0000:10:00.0/firmware/0000:10:00.0/loading
[   31.298572] CPU 1
[   31.298574] Modules linked in: iscsi_tcp libiscsi_tcp libiscsi
scsi_transport_iscsi sbp2 btusb sdhci_pci sdhci bluetooth psmouse
ricoh_mmc ohci1394 mmc_core ieee1394 pl2303
[   31.298585] Pid: 0, comm: swapper Not tainted 2.6.30 #9 HP Compaq
6910p (GB949ET#AK8)
[   31.298587] RIP: c2c0:[<ffffffff814f6940>]  [<ffffffff814f6940>]
fair_sched_class+0x0/0xc0
[   31.298592] RSP: ffffffff8102e891:ffff88013b9a1b20  EFLAGS:
ffff88013b8c4a80
[   31.298594] RAX: ffffffff81030f1a RBX: ffffffff810b047b RCX:
ffffffff8123d135
[   31.298597] RDX: ffff88013b976580 RSI: 0000000000000001 RDI:
ffff88013b8c4a80
[   31.298599] RBP: ffffffff810303a2 R08: ffff88013b9a1ae0 R09:
ffff88002804c2c0
[   31.298601] R10: ffff88013b9765b8 R11: ffff88002804c328 R12:
ffff88013b9a1ab0
[   31.298603] R13: 0000000000000001 R14: 0000000000000001 R15:
ffff88002804c328
[   31.298606] FS:  000000[   31.319076] bad: scheduling from the idle
thread!
[   31.319[   31.320577] Call Trace:
[   31.321325] bad: scheduling from the idle thread!
[   31.322044] Pid: 0, comm: swapper Tainted: [   31.322805] Call Trace:
[   31.323568] bad: scheduling from the idle thread!
[   31.324318] Pid: 0, comm: swapper Tainted: G      D    2.6.30 #9
[   31.325028] Call Trace:

I did a few runs and the full output from that is here:
http://pastebin.com/fe18c31c

fair_sched_class seems to be mentioned every time from what I've seen.

I also tried removing "button" support from ACPI and that seems to
eliminate the problem, so I have a workaround.
Compiling "button" as a module does not seem to eliminate the problem
though even with the module blacklisted (and not loaded then of course).

This is on a HP Compaq 6910p as you might have seen in the oops/panic
output.
I don't know how to proceed on troubleshooting/debugging this.

Let me know if there is anything I can do to help find the issue.

P.S. I'm currently not subscribed so please feel free to CC me on any
posts regarding
this.

/Andreas Sundstrom

Re: 2.6.30 kernel panic when closing lid

[Frans Pop]

Andreas Sundstrom wrote:
> My laptop has started crashing when I close the lid. This does not
> happen with 2.6.29.4 which I was using before 2.6.30, I configured my
> 2.6.30 kernel with the 2.6.29.4 as the base and then did "make
> oldconfig" and then most of the questions I got I used the default
> setting.

That sounds like it may be a known issue, although it's somewhat strange 
that you did not see it with earlier kernels.
Can you test if the following workaround fixes the crash?

echo -n 7 >/proc/acpi/video/C09A/DOS

See e.g.: http://lkml.indiana.edu/hypermail/linux/kernel/0901.1/01971.html
You can find more information by googling for "hp 2510p lid close crash"
(but it affects other models as well).

Cheers,
FJP

Re: 2.6.30 kernel panic when closing lid

[Andreas Sundstrom]

Frans Pop wrote:
> Andreas Sundstrom wrote:
>> My laptop has started crashing when I close the lid. This does not
>> happen with 2.6.29.4 which I was using before 2.6.30, I configured my
>> 2.6.30 kernel with the 2.6.29.4 as the base and then did "make
>> oldconfig" and then most of the questions I got I used the default
>> setting.
> 
> That sounds like it may be a known issue, although it's somewhat strange 
> that you did not see it with earlier kernels.
> Can you test if the following workaround fixes the crash?
> 
> echo -n 7 >/proc/acpi/video/C09A/DOS
> 
> See e.g.: http://lkml.indiana.edu/hypermail/linux/kernel/0901.1/01971.html
> You can find more information by googling for "hp 2510p lid close crash"
> (but it affects other models as well).
> 
> Cheers,
> FJP

I think you hit the nail on the head there.. It seems to work now with
DOS set to 7 instead of 0 as it defaults to.

Funny thing is that I know remember going through this issue some time
ago.. I have since then changed from ubuntu to debian lenny. It used to
be taken cared of by a hotkeys script in init.d and I have now installed
 the corresponding debian package for that and all is well.

Many thanks for your help, funny how fast one can forget sometimes..
Strange though that I did not have any issues with 2.6.28, 2.6.29 (not
that I can remember at least).

/Andreas

Re: 2.6.30 kernel panic when closing lid

[Bjorn Helgaas]

On Friday 19 June 2009 02:22:00 pm Andreas Sundstrom wrote:
> My laptop has started crashing when I close the lid. This does not
> happen with 2.6.29.4 which I was using before 2.6.30, ...

I opened a bugzilla for this regression:
  http://bugzilla.kernel.org/show_bug.cgi?id=13751

If either of you can test the suggestions in comment #12 of the
bugzilla, I'd like to try to fix this oops.

Thanks!
  Bjorn

> I have tried to catch the full output with a USB serial console, but
> it's not able to send the full output before it dies it seems. Sometimes
> it gets further than other times.
> 
> Here is one oops/panic:
> [   31.295898] kernel tried to execute NX-protected page - exploit
> attempt? (uid: 0)
> [   31.296575] BUG: unable to handle kernel paging request at
> ffff88013a891630
> [   31.297233] IP: [<ffffffff814f6940>] fair_sched_class+0x0/0xc0
> [   31.297594] PGD 202063 PUD b067 PMD 800000013a8001e3
> [   31.298567] Oops: 0011 [#1] SMP
> [   31.298569] last sysfs file:
> /sys/devices/pci0000:00/0000:00:1c.1/0000:10:00.0/firmware/0000:10:00.0/loading
> [   31.298572] CPU 1
> [   31.298574] Modules linked in: iscsi_tcp libiscsi_tcp libiscsi
> scsi_transport_iscsi sbp2 btusb sdhci_pci sdhci bluetooth psmouse
> ricoh_mmc ohci1394 mmc_core ieee1394 pl2303
> [   31.298585] Pid: 0, comm: swapper Not tainted 2.6.30 #9 HP Compaq
> 6910p (GB949ET#AK8)
> [   31.298587] RIP: c2c0:[<ffffffff814f6940>]  [<ffffffff814f6940>]
> fair_sched_class+0x0/0xc0
> [   31.298592] RSP: ffffffff8102e891:ffff88013b9a1b20  EFLAGS:
> ffff88013b8c4a80
> [   31.298594] RAX: ffffffff81030f1a RBX: ffffffff810b047b RCX:
> ffffffff8123d135
> [   31.298597] RDX: ffff88013b976580 RSI: 0000000000000001 RDI:
> ffff88013b8c4a80
> [   31.298599] RBP: ffffffff810303a2 R08: ffff88013b9a1ae0 R09:
> ffff88002804c2c0
> [   31.298601] R10: ffff88013b9765b8 R11: ffff88002804c328 R12:
> ffff88013b9a1ab0
> [   31.298603] R13: 0000000000000001 R14: 0000000000000001 R15:
> ffff88002804c328
> [   31.298606] FS:  000000[   31.319076] bad: scheduling from the idle
> thread!
> [   31.319[   31.320577] Call Trace:
> [   31.321325] bad: scheduling from the idle thread!
> [   31.322044] Pid: 0, comm: swapper Tainted: [   31.322805] Call Trace:
> [   31.323568] bad: scheduling from the idle thread!
> [   31.324318] Pid: 0, comm: swapper Tainted: G      D    2.6.30 #9
> [   31.325028] Call Trace:
> 
> I did a few runs and the full output from that is here:
> http://pastebin.com/fe18c31c
> 
> fair_sched_class seems to be mentioned every time from what I've seen.
> 
> I also tried removing "button" support from ACPI and that seems to
> eliminate the problem, so I have a workaround.
> Compiling "button" as a module does not seem to eliminate the problem
> though even with the module blacklisted (and not loaded then of course).
> 
> This is on a HP Compaq 6910p as you might have seen in the oops/panic
> output.
> I don't know how to proceed on troubleshooting/debugging this.
> 
> Let me know if there is anything I can do to help find the issue.
> 
> P.S. I'm currently not subscribed so please feel free to CC me on any
> posts regarding
> this.
> 
> /Andreas Sundstrom