LinuxLists.cc - 2.6.24-rc1: OOPS at acpi_battery

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

On Mon, 29 Oct 2007 11:11:04 +0100
Romano Giannetti <[email protected]> wrote:

>
> Hi,
>
> sometime on resuming from s2ram my laptop spew the following oops.
> Config, dmesg etc are at:
>
> http://www.dea.icai.upcomillas.es/romano/linux/info/2624rc1_6/
>
> [ 3.475386] Oops: 0000 [#1] SMP
> [ 3.475602] Process kacpi_notify (pid: 50, ti=c2122000 task=c210c030 task.ti=c2122000)
> [ 3.475608] Stack: c35c5060 c02f2a58 00000046 00000000 6b6b6b6b c2b20298 c2123efc c02f29ac
> [ 3.475626] c02166cb f896f3ca 00000000 f896ff63 00000246 00000000 c2b20298 ffffa512
> [ 3.475644] c2123f04 c02f2a58 c2123f34 f896f3ed c2123f1c c34fbdb8 c2123f20 c02f3e22
> [ 3.475662] Call Trace:
> [ 3.475667] [show_trace_log_lvl+26/48] show_trace_log_lvl+0x1a/0x30
> [ 3.475678] [show_stack_log_lvl+177/224] show_stack_log_lvl+0xb1/0xe0
> [ 3.475695] [die+282/560] die+0x11a/0x230
> [ 3.475687] [show_registers+193/464] show_registers+0xc1/0x1d0
> [ 3.475703] [do_page_fault+415/1648] do_page_fault+0x19f/0x670
> [ 3.475713] [error_code+114/120] error_code+0x72/0x78
> [ 3.475722] [__mutex_unlock_slowpath+172/336] __mutex_unlock_slowpath+0xac/0x150
> [ 3.475731] [mutex_unlock+8/16] mutex_unlock+0x8/0x10
> [ 3.475739] [<f896f3ed>] acpi_battery_update+0x1ce/0x23c [battery]
> [ 3.475753] [<f896f927>] acpi_battery_notify+0x21/0x78 [battery]
> [ 3.475764] [acpi_ev_notify_dispatch+79/90] acpi_ev_notify_dispatch+0x4f/0x5a
> [ 3.475792] [worker_thread+157/256] worker_thread+0x9d/0x100
> [ 3.475774] [acpi_os_execute_notify+36/47] acpi_os_execute_notify+0x24/0x2f
> [ 3.475784] [run_workqueue+288/464] run_workqueue+0x120/0x1d0
> [ 3.475809] [kernel_thread_helper+7/16] kernel_thread_helper+0x7/0x10
> [ 3.475801] [kthread+66/112] kthread+0x42/0x70
> [ 3.475821] Code: 8d b4 26 00 00 00 00 55 89 e5 83 ec 18 89 5d f8 89 c3 89 75 fc 0f b6 40 04 89 d6 84 c0 7f 24 8d 43 20 3b 43 20 0f 84 e4 00 00 00 <3b> 76 10 0f 85 93 00 00 00 3b 36 90 74 47 8b 5d f8 8b 75 fc 89
> [ 3.475818] =======================
> [ 3.475909] EIP: [debug_mutex_wake_waiter+36/352] debug_mutex_wake_waiter+0x24/0x160 SS:ESP 0068:c2123ebc
>

Did any earlier kernels do this? In other words, do you believe that this
is a bug which we added after 2.6.23 was released?

Thanks.

2007-11-02 15:52:19

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

On Friday, 2 November 2007 00:14, Andrew Morton wrote:
> On Mon, 29 Oct 2007 11:11:04 +0100
> Romano Giannetti <[email protected]> wrote:
>
> >
> > Hi,
> >
> > sometime on resuming from s2ram my laptop spew the following oops.
> > Config, dmesg etc are at:
> >
> > http://www.dea.icai.upcomillas.es/romano/linux/info/2624rc1_6/
> >
> > [ 3.475386] Oops: 0000 [#1] SMP
> > [ 3.475602] Process kacpi_notify (pid: 50, ti=c2122000 task=c210c030 task.ti=c2122000)
> > [ 3.475608] Stack: c35c5060 c02f2a58 00000046 00000000 6b6b6b6b c2b20298 c2123efc c02f29ac
> > [ 3.475626] c02166cb f896f3ca 00000000 f896ff63 00000246 00000000 c2b20298 ffffa512
> > [ 3.475644] c2123f04 c02f2a58 c2123f34 f896f3ed c2123f1c c34fbdb8 c2123f20 c02f3e22
> > [ 3.475662] Call Trace:
> > [ 3.475667] [show_trace_log_lvl+26/48] show_trace_log_lvl+0x1a/0x30
> > [ 3.475678] [show_stack_log_lvl+177/224] show_stack_log_lvl+0xb1/0xe0
> > [ 3.475695] [die+282/560] die+0x11a/0x230
> > [ 3.475687] [show_registers+193/464] show_registers+0xc1/0x1d0
> > [ 3.475703] [do_page_fault+415/1648] do_page_fault+0x19f/0x670
> > [ 3.475713] [error_code+114/120] error_code+0x72/0x78
> > [ 3.475722] [__mutex_unlock_slowpath+172/336] __mutex_unlock_slowpath+0xac/0x150
> > [ 3.475731] [mutex_unlock+8/16] mutex_unlock+0x8/0x10
> > [ 3.475739] [<f896f3ed>] acpi_battery_update+0x1ce/0x23c [battery]
> > [ 3.475753] [<f896f927>] acpi_battery_notify+0x21/0x78 [battery]
> > [ 3.475764] [acpi_ev_notify_dispatch+79/90] acpi_ev_notify_dispatch+0x4f/0x5a
> > [ 3.475792] [worker_thread+157/256] worker_thread+0x9d/0x100
> > [ 3.475774] [acpi_os_execute_notify+36/47] acpi_os_execute_notify+0x24/0x2f
> > [ 3.475784] [run_workqueue+288/464] run_workqueue+0x120/0x1d0
> > [ 3.475809] [kernel_thread_helper+7/16] kernel_thread_helper+0x7/0x10
> > [ 3.475801] [kthread+66/112] kthread+0x42/0x70
> > [ 3.475821] Code: 8d b4 26 00 00 00 00 55 89 e5 83 ec 18 89 5d f8 89 c3 89 75 fc 0f b6 40 04 89 d6 84 c0 7f 24 8d 43 20 3b 43 20 0f 84 e4 00 00 00 <3b> 76 10 0f 85 93 00 00 00 3b 36 90 74 47 8b 5d f8 8b 75 fc 89
> > [ 3.475818] =======================
> > [ 3.475909] EIP: [debug_mutex_wake_waiter+36/352] debug_mutex_wake_waiter+0x24/0x160 SS:ESP 0068:c2123ebc
> >
>
> Did any earlier kernels do this? In other words, do you believe that this
> is a bug which we added after 2.6.23 was released?

This might be the same as http://bugzilla.kernel.org/show_bug.cgi?id=9283

2007-11-04 09:34:26

by Romano Giannetti

[permalink] [raw]

Subject: Re: SPAM Re: 2.6.24-rc1: OOPS at acpi_battery_update

On Fri, 2007-11-02 at 17:08 +0100, Rafael J. Wysocki wrote:
> On Friday, 2 November 2007 00:14, Andrew Morton wrote:
> > On Mon, 29 Oct 2007 11:11:04 +0100
[OOPS removed]
> >
> > Did any earlier kernels do this? In other words, do you believe that this
> > is a bug which we added after 2.6.23 was released?

Never noticed it before. But it's happening just sometime (I've had it
just two times), so I cannot be sure.

> This might be the same as http://bugzilla.kernel.org/show_bug.cgi?id=9283

Hmmm... there is too little info for my level in that bug report to
understand if it's the same.

Romano

--
Sorry for the disclaimer --- ?I cannot stop it!

--
La presente comunicaci?n tiene car?cter confidencial y es para el exclusivo uso del destinatario indicado en la misma. Si Ud. no es el destinatario indicado, le informamos que cualquier forma de distribuci?n, reproducci?n o uso de esta comunicaci?n y/o de la informaci?n contenida en la misma est?n estrictamente prohibidos por la ley. Si Ud. ha recibido esta comunicaci?n por error, por favor, notif?quelo inmediatamente al remitente contestando a este mensaje y proceda a continuaci?n a destruirlo. Gracias por su colaboraci?n.

This communication contains confidential information. It is for the exclusive use of the intended addressee. If you are not the intended addressee, please note that any form of distribution, copying or use of this communication or the information in it is strictly prohibited by law. If you have received this communication in error, please immediately notify the sender by reply e-mail and destroy this message. Thank you for your cooperation.

2007-11-04 13:00:33

by Michael (rabenkind) Brandstetter

[permalink] [raw]

Subject: Re: SPAM Re: 2.6.24-rc1: OOPS at acpi_battery_update

On Sunday, 4 of November 2007, Romano Giannetti wrote:
>
> On Fri, 2007-11-02 at 17:08 +0100, Rafael J. Wysocki wrote:
> > On Friday, 2 November 2007 00:14, Andrew Morton wrote:
> > > On Mon, 29 Oct 2007 11:11:04 +0100
> [OOPS removed]
> > >
> > > Did any earlier kernels do this? In other words, do you believe that this
> > > is a bug which we added after 2.6.23 was released?
>
> Never noticed it before. But it's happening just sometime (I've had it
> just two times), so I cannot be sure.
>
> > This might be the same as http://bugzilla.kernel.org/show_bug.cgi?id=9283
>
> Hmmm... there is too little info for my level in that bug report to
> understand if it's the same.

Then, please apply the patch from

http://bugzilla.kernel.org/attachment.cgi?id=13376&action=view

and see if the problem goes away.

2007-11-04 23:36:20

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

Am Sonntag, 4. November 2007 14:17:23 schrieben Sie:
> On Sunday, 4 of November 2007, Romano Giannetti wrote:
> > On Fri, 2007-11-02 at 17:08 +0100, Rafael J. Wysocki wrote:
> > > On Friday, 2 November 2007 00:14, Andrew Morton wrote:
> > > > On Mon, 29 Oct 2007 11:11:04 +0100

Hello!

> > [OOPS removed]
> >
> > > > Did any earlier kernels do this? In other words, do you believe that
> > > > this is a bug which we added after 2.6.23 was released?

Yes the kernels prior to 2.6.24-rc1 did not have this bug.

The oops always appear when i plug in or remove the battery.

> > Never noticed it before. But it's happening just sometime (I've had it
> > just two times), so I cannot be sure.
> >
> > > This might be the same as
> > > http://bugzilla.kernel.org/show_bug.cgi?id=9283
> >
> > Hmmm... there is too little info for my level in that bug report to
> > understand if it's the same.
>
> Then, please apply the patch from
>
> http://bugzilla.kernel.org/attachment.cgi?id=13376&action=view
>
> and see if the problem goes away.

The bug went away on my Dell Latitude C610. I applied the patch on two boxes
(2.6.24-rc1-git12) and the haldaemon is not killed and no oops when i remove
the battery or plug it in.

But on the Portege 3110CT the bug remains.

greetz Michael :))
--
Duisburger Linux User Group
Free Software Foundation Europe

/"\ ASCII Ribbon Campaign
\ / No HTML/RTF in email
x No Word docs in email
/ \ Respect for open standards

2007-11-05 00:01:44

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

On Monday, 5 of November 2007, Michael (rabenkind) Brandstetter wrote:
> Am Sonntag, 4. November 2007 14:17:23 schrieben Sie:
> > On Sunday, 4 of November 2007, Romano Giannetti wrote:
> > > On Fri, 2007-11-02 at 17:08 +0100, Rafael J. Wysocki wrote:
> > > > On Friday, 2 November 2007 00:14, Andrew Morton wrote:
> > > > > On Mon, 29 Oct 2007 11:11:04 +0100
>
> Hello!
>
> > > [OOPS removed]
> > >
> > > > > Did any earlier kernels do this? In other words, do you believe that
> > > > > this is a bug which we added after 2.6.23 was released?
>
> Yes the kernels prior to 2.6.24-rc1 did not have this bug.
>
> The oops always appear when i plug in or remove the battery.
>
> > > Never noticed it before. But it's happening just sometime (I've had it
> > > just two times), so I cannot be sure.
> > >
> > > > This might be the same as
> > > > http://bugzilla.kernel.org/show_bug.cgi?id=9283
> > >
> > > Hmmm... there is too little info for my level in that bug report to
> > > understand if it's the same.
> >
> > Then, please apply the patch from
> >
> > http://bugzilla.kernel.org/attachment.cgi?id=13376&action=view
> >
> > and see if the problem goes away.
>
> The bug went away on my Dell Latitude C610. I applied the patch on two boxes
> (2.6.24-rc1-git12) and the haldaemon is not killed and no oops when i remove
> the battery or plug it in.
>
> But on the Portege 3110CT the bug remains.

Well, please create a bug report at http://bugzilla.kernel.org regarding this
bug and place all of the relevant information in there (including what boxes
are fixed by the above patch etc.) Also please add my address to the report's
CC list.

Thanks,
Rafael

2007-11-08 15:55:21

by Johannes Weiner

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

Hi,

is there any reason, why acpi_battery_get_property() should call
acpi_battery_update() at all?

Hannes

2007-11-08 16:00:07

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

On Thursday, 8 of November 2007, Johannes Weiner wrote:
> Hi,
>
> is there any reason, why acpi_battery_get_property() should call
> acpi_battery_update() at all?

Alex?

2007-11-08 16:12:18

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

Rafael J. Wysocki wrote:
> On Thursday, 8 of November 2007, Johannes Weiner wrote:
>> Hi,
>>
>> is there any reason, why acpi_battery_get_property() should call
>> acpi_battery_update() at all?
>
> Alex?
If someone wants to read stale values, he could comment out acpi_battery_update.

Regards,
Alex.

2007-11-08 16:35:54

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

Rafael J. Wysocki wrote:
> On Thursday, 8 of November 2007, Johannes Weiner wrote:
>> Hi,
>>
>> is there any reason, why acpi_battery_get_property() should call
>> acpi_battery_update() at all?
>
> Alex?
Do you mean "why should it call _whole_ battery update?" ?
get_state should be sufficient in order to not get stale data.

Regards,
Alex.

Attachments:

remove_cycle_at_battery_removal.patch (1.69 kB)

2007-11-08 16:47:38

by Johannes Weiner

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

Hi,

On Thu, Nov 08, 2007 at 07:11:53PM +0300, Alexey Starikovskiy wrote:
> Rafael J. Wysocki wrote:
> >On Thursday, 8 of November 2007, Johannes Weiner wrote:
> >>Hi,
> >>
> >>is there any reason, why acpi_battery_get_property() should call
> >>acpi_battery_update() at all?
> >
> >Alex?
> If someone wants to read stale values, he could comment out
> acpi_battery_update.

Please help me to understand:

When the battery is plugged in, the acpi_battery_notify() is called,
which in turn calls acpi_battery_update(). The latter ensures that the
sysfs files are created if not yet present.

When the battery is removed, acpi_battery_notify is called, which in
turn calls acpi_battery_update(). The latter ensures that the sysfs
files are removed if present.

During runtime - as far as I understood - no sysfs files have to be
created/removed but the saved battery state info becomes stale.

So, would it be enough to call acpi_battery_get_state() in
acpi_battery_get_property() instead of acpi_battery_update()?

If acpi_battery_get_state() does what I think it does, this would
ensure that the battery state is reread and updated when
acpi_battery_get_property() is called.

Hit me.

Hannes

2007-11-08 17:00:32

by Johannes Weiner

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

Hi Alex,

On Thu, Nov 08, 2007 at 07:35:23PM +0300, Alexey Starikovskiy wrote:
> Rafael J. Wysocki wrote:
> >On Thursday, 8 of November 2007, Johannes Weiner wrote:
> >>Hi,
> >>
> >>is there any reason, why acpi_battery_get_property() should call
> >>acpi_battery_update() at all?
> >
> >Alex?
> Do you mean "why should it call _whole_ battery update?" ?
> get_state should be sufficient in order to not get stale data.

Okay, than I understood it correct.

Acked-by: Johannes Weiner <[email protected]>

Thanks!

Hannes

2007-11-09 04:35:47

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

A> On Thu, 08 Nov 2007 19:35:23 +0300 Alexey Starikovskiy <[email protected]> wrote:
> [remove_cycle_at_battery_removal.patch text/x-patch (1.7KB)]
> ACPI: Battery: remove cycle from battery removal.
>
> From: Alexey Starikovskiy <[email protected]>
>
> get_property() should not call battery_update() on absent battery to
> avoid cycle and oops.
>
> Signed-off-by: Alexey Starikovskiy <[email protected]>
> Tested-by: Rolf Eike Beer <[email protected]>

A patch similar to this one but with an identical changelog was merged into
Len's tree on November 2.

If it had been promptly merged into mainline then quite a lot of people's
time would not have been wasted.

2007-11-09 09:37:21

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

Andrew Morton wrote:
> A> On Thu, 08 Nov 2007 19:35:23 +0300 Alexey Starikovskiy <[email protected]> wrote:
>> [remove_cycle_at_battery_removal.patch text/x-patch (1.7KB)]
>> ACPI: Battery: remove cycle from battery removal.
>>
>> From: Alexey Starikovskiy <[email protected]>
>>
>> get_property() should not call battery_update() on absent battery to
>> avoid cycle and oops.
>>
>> Signed-off-by: Alexey Starikovskiy <[email protected]>
>> Tested-by: Rolf Eike Beer <[email protected]>
>
> A patch similar to this one but with an identical changelog was merged into
> Len's tree on November 2.
>
> If it had been promptly merged into mainline then quite a lot of people's
> time would not have been wasted.
>
Andrew,
should I send such patches directly to you/Linus?

Regards,
Alex

2007-11-09 09:50:42

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

On Fri, 09 Nov 2007 12:36:43 +0300 Alexey Starikovskiy <[email protected]> wrote:

> Andrew Morton wrote:
> > A> On Thu, 08 Nov 2007 19:35:23 +0300 Alexey Starikovskiy <[email protected]> wrote:
> >> [remove_cycle_at_battery_removal.patch text/x-patch (1.7KB)]
> >> ACPI: Battery: remove cycle from battery removal.
> >>
> >> From: Alexey Starikovskiy <[email protected]>
> >>
> >> get_property() should not call battery_update() on absent battery to
> >> avoid cycle and oops.
> >>
> >> Signed-off-by: Alexey Starikovskiy <[email protected]>
> >> Tested-by: Rolf Eike Beer <[email protected]>
> >
> > A patch similar to this one but with an identical changelog was merged into
> > Len's tree on November 2.
> >
> > If it had been promptly merged into mainline then quite a lot of people's
> > time would not have been wasted.
> >
> Andrew,
> should I send such patches directly to you/Linus?
>

Well if Len doesn't object and you're confident in the changes, why not?
Any time we leave bugs unfixed we drive away testers and that impacts all
parts of the kernel.

2007-11-13 08:35:35

[permalink] [raw]

Subject: Re: 2.6.24-rc1: OOPS at acpi_battery_update

Andrew,

I can not contact with Len for several days, while the oops on battery
seems quite important.
It also seem to behave well in -mm tree (as part of Len's acpi-test).
Will you send this patch to Linus without approval from Len or should I?

Thanks,
Alex.

Andrew Morton wrote:
> On Fri, 09 Nov 2007 12:36:43 +0300 Alexey Starikovskiy <[email protected]> wrote:
>
>
>> Andrew Morton wrote:
>>
>>> A> On Thu, 08 Nov 2007 19:35:23 +0300 Alexey Starikovskiy <[email protected]> wrote:
>>>
>>>> [remove_cycle_at_battery_removal.patch text/x-patch (1.7KB)]
>>>> ACPI: Battery: remove cycle from battery removal.
>>>>
>>>> From: Alexey Starikovskiy <[email protected]>
>>>>
>>>> get_property() should not call battery_update() on absent battery to
>>>> avoid cycle and oops.
>>>>
>>>> Signed-off-by: Alexey Starikovskiy <[email protected]>
>>>> Tested-by: Rolf Eike Beer <[email protected]>
>>>>
>>> A patch similar to this one but with an identical changelog was merged into
>>> Len's tree on November 2.
>>>
>>> If it had been promptly merged into mainline then quite a lot of people's
>>> time would not have been wasted.
>>>
>>>
>> Andrew,
>> should I send such patches directly to you/Linus?
>>
>>
>
> Well if Len doesn't object and you're confident in the changes, why not?
> Any time we leave bugs unfixed we drive away testers and that impacts all
> parts of the kernel.
>
>

Attachments:

remove_cycle_at_battery_removal.patch (1.69 kB)

2007-11-13 09:00:16