Linux 2.6.34 causes system lockup on Compaq Presario 2200 Laptop

Linux 2.6.34 causes system lockup on Compaq Presario 2200 Laptop

[Jeffrey Merkey]

2.6.34 causes a hard system lockup when the screen blanking activates
for more than 10 minutes with low battery power.  Also misreports
battery level as "battery broken at 14%" when the battery is fully
charged.  Does not occur with 2.6.28 but 2.6.33 also seems to do it as
well.  software debugging tools useless as the keyboard does not
respond.

Jeff

Re: Linux 2.6.34 causes system lockup on Compaq Presario 2200 Laptop

[Andrew Morton]

(cc linux-acpi)

On Thu, 27 May 2010 12:08:48 -0600 Jeffrey Merkey <jeffmerkey@gmail.com> wrote:

> 2.6.34 causes a hard system lockup when the screen blanking activates
> for more than 10 minutes with low battery power.  Also misreports
> battery level as "battery broken at 14%" when the battery is fully
> charged.  Does not occur with 2.6.28 but 2.6.33 also seems to do it as
> well.  software debugging tools useless as the keyboard does not
> respond.
> 

Re: Linux 2.6.34 causes system lockup on Compaq Presario 2200 Laptop

[Len Brown]

Jeff,
Please attach the output of acpidump here:

https://bugzilla.kernel.org/show_bug.cgi?id=16136

thanks,
Len Brown, Intel Open Source Technology Center

Re: Linux 2.6.34 causes system lockup on Compaq Presario 2200 Laptop

[Jeffrey Merkey]

Len,

I am having trouble getting a trace of this because its down hard to
reproduce.  The reason is the lockup happens before I log into the
laptop.  i.e. I boot 2.6.34 and if I let the laptop just sit there is
happens.  If I log in and start doing work on it, it does not seem to
happen.  It's related to using the laptop on battery mode with the
charger removed.  I have a 2U appliance setup I am working on today
and this frees up the laptop to try to reproduce this.  The flaky
battery power misreporting was in fact accurate.  The battery in that
laptop is 6 years old and does have low power.  I am certain the
lockup if related though.  I will try to reproduce today and setup the
trace to run after boot.  I have had it sitting there for an  hour and
it has not reoccurred.  Let me try through today and if I cannot
reproduce, close the bug.

Jeff

On Wed, Jun 9, 2010 at 4:21 PM, Len Brown <lenb@kernel.org> wrote:
> Jeff,
> Please attach the output of acpidump here:
>
> https://bugzilla.kernel.org/show_bug.cgi?id=16136
>
> thanks,
> Len Brown, Intel Open Source Technology Center
>

Re: Linux 2.6.34 causes system lockup on Compaq Presario 2200 Laptop

[Len Brown]

On Fri, 11 Jun 2010, Jeffrey Merkey wrote:

> Len,
> 
> I am having trouble getting a trace of this because its down hard to
> reproduce.  The reason is the lockup happens before I log into the
> laptop.  i.e. I boot 2.6.34 and if I let the laptop just sit there is
> happens.  If I log in and start doing work on it, it does not seem to
> happen.  It's related to using the laptop on battery mode with the
> charger removed.  I have a 2U appliance setup I am working on today
> and this frees up the laptop to try to reproduce this.  The flaky
> battery power misreporting was in fact accurate.  The battery in that
> laptop is 6 years old and does have low power.  I am certain the
> lockup if related though.  I will try to reproduce today and setup the
> trace to run after boot.  I have had it sitting there for an  hour and
> it has not reoccurred.  Let me try through today and if I cannot
> reproduce, close the bug.

Good to know, Jeff.

One possibility is that the old battery is confusing
the embedded controller -- possibly exposing a firmware bug.
The EC, in turn, is fully capable of hanging the OS.

If you *can* reproduce it, it may be interesting to bisect
recent changes to drivers/acpi/ec.c

Also, if you can reproduce it, it would be good to verify
that the problem occurs only when the battery driver is loaded.

thanks,
-Len

> > Please attach the output of acpidump here:
> >
> > https://bugzilla.kernel.org/show_bug.cgi?id=16136

We could still use this output -- note that you can
capture it at any time, independent of the failure,
and independent of the kernel -- as it is simply
a copy of the tables that the BIOS is exporting to Linux.

thanks,
-Len Brown, Intel Open Source Technology Center

Re: Linux 2.6.34 causes system lockup on Compaq Presario 2200 Laptop

[Jeffrey Merkey]

On Fri, Jun 11, 2010 at 11:07 AM, Len Brown <lenb@kernel.org> wrote:
> On Fri, 11 Jun 2010, Jeffrey Merkey wrote:
>
>> Len,
>>
>> I am having trouble getting a trace of this because its down hard to
>> reproduce.  The reason is the lockup happens before I log into the
>> laptop.  i.e. I boot 2.6.34 and if I let the laptop just sit there is
>> happens.  If I log in and start doing work on it, it does not seem to
>> happen.  It's related to using the laptop on battery mode with the
>> charger removed.  I have a 2U appliance setup I am working on today
>> and this frees up the laptop to try to reproduce this.  The flaky
>> battery power misreporting was in fact accurate.  The battery in that
>> laptop is 6 years old and does have low power.  I am certain the
>> lockup if related though.  I will try to reproduce today and setup the
>> trace to run after boot.  I have had it sitting there for an  hour and
>> it has not reoccurred.  Let me try through today and if I cannot
>> reproduce, close the bug.
>
> Good to know, Jeff.
>
> One possibility is that the old battery is confusing
> the embedded controller -- possibly exposing a firmware bug.
> The EC, in turn, is fully capable of hanging the OS.
>
> If you *can* reproduce it, it may be interesting to bisect
> recent changes to drivers/acpi/ec.c
>
> Also, if you can reproduce it, it would be good to verify
> that the problem occurs only when the battery driver is loaded.
>
> thanks,
> -Len
>
>> > Please attach the output of acpidump here:
>> >
>> > https://bugzilla.kernel.org/show_bug.cgi?id=16136
>
> We could still use this output -- note that you can
> capture it at any time, independent of the failure,
> and independent of the kernel -- as it is simply
> a copy of the tables that the BIOS is exporting to Linux.
>
> thanks,
> -Len Brown, Intel Open Source Technology Center
>

I can tell you for certain it does only happen when the battery driver
is loaded.  Still has not ocurred.  Just rebooted it 15 minutes ago --
trying again.

Jeff

Re: Linux 2.6.34 causes system lockup on Compaq Presario 2200 Laptop

[Jeffrey Merkey]

Not Reliably Reproducable.  Looks like a firmware/hardware problem
with an old laptop.

Jeff

On Fri, Jun 11, 2010 at 3:19 PM, Jeffrey Merkey <jeffmerkey@gmail.com> wrote:
> On Fri, Jun 11, 2010 at 11:07 AM, Len Brown <lenb@kernel.org> wrote:
>> On Fri, 11 Jun 2010, Jeffrey Merkey wrote:
>>
>>> Len,
>>>
>>> I am having trouble getting a trace of this because its down hard to
>>> reproduce.  The reason is the lockup happens before I log into the
>>> laptop.  i.e. I boot 2.6.34 and if I let the laptop just sit there is
>>> happens.  If I log in and start doing work on it, it does not seem to
>>> happen.  It's related to using the laptop on battery mode with the
>>> charger removed.  I have a 2U appliance setup I am working on today
>>> and this frees up the laptop to try to reproduce this.  The flaky
>>> battery power misreporting was in fact accurate.  The battery in that
>>> laptop is 6 years old and does have low power.  I am certain the
>>> lockup if related though.  I will try to reproduce today and setup the
>>> trace to run after boot.  I have had it sitting there for an  hour and
>>> it has not reoccurred.  Let me try through today and if I cannot
>>> reproduce, close the bug.
>>
>> Good to know, Jeff.
>>
>> One possibility is that the old battery is confusing
>> the embedded controller -- possibly exposing a firmware bug.
>> The EC, in turn, is fully capable of hanging the OS.
>>
>> If you *can* reproduce it, it may be interesting to bisect
>> recent changes to drivers/acpi/ec.c
>>
>> Also, if you can reproduce it, it would be good to verify
>> that the problem occurs only when the battery driver is loaded.
>>
>> thanks,
>> -Len
>>
>>> > Please attach the output of acpidump here:
>>> >
>>> > https://bugzilla.kernel.org/show_bug.cgi?id=16136
>>
>> We could still use this output -- note that you can
>> capture it at any time, independent of the failure,
>> and independent of the kernel -- as it is simply
>> a copy of the tables that the BIOS is exporting to Linux.
>>
>> thanks,
>> -Len Brown, Intel Open Source Technology Center
>>
>
> I can tell you for certain it does only happen when the battery driver
> is loaded.  Still has not ocurred.  Just rebooted it 15 minutes ago --
> trying again.
>
> Jeff
>

Re: Linux 2.6.34 causes system lockup on Compaq Presario 2200 Laptop

[Maciej Rutecki]

On czwartek, 27 maja 2010 o 20:08:48 Jeffrey Merkey wrote:
> 2.6.34 causes a hard system lockup when the screen blanking activates
> for more than 10 minutes with low battery power.  Also misreports
> battery level as "battery broken at 14%" when the battery is fully
> charged.  Does not occur with 2.6.28 but 2.6.33 also seems to do it as
> well.  software debugging tools useless as the keyboard does not
> respond.
> 
> Jeff

I created a Bugzilla entry at 
https://bugzilla.kernel.org/show_bug.cgi?id=16136
for your bug report, please add your address to the CC list in there, thanks!

-- 
Maciej Rutecki
http://www.maciek.unixy.pl