3.1 Panic on power event

3.1 Panic on power event

[Josh Boyer]

Hi All,

We've had a report [1] from a user that the 3.1.6 kernel panics if they
plug or unplug their machine.  The oops seems to come from an issue
somewhere in the acpi notification stack.  The relevant backtrace
functions are (transcribed from a jpeg):

acpi_ac_notify
acpi_device_notify
acpi_ev_notify_dispatch
acpi_os_execute_deferred

Looking at the dmesg during boot, it seems this machine is also
triggering an oops for duplicate registration of an ACAD device.  You
can find the backtraces for that below.

Looking at the acpidump output [2], I noticed that there is indeed two
Device(ACAD) entries in the DSDT.  The first is under the system bus and
the second is under PCI0.  The problem seems to stem from some
Notify(ACAD) entries, but ACAD is ambiguous.  I'm guessing the ACPI code
is getting an event for the device that wasn't set up properly?

I'm by no means an ACPI expert, so I'd really appreciate thoughts and
review.

josh

[1] https://bugzilla.redhat.com/show_bug.cgi?id=772730
[2] https://bugzilla.redhat.com/attachment.cgi?id=551677

Duplicate device oopses:

ACPI: Deprecated procfs I/F for AC is loaded, please retry with CONFIG_ACPI_PROCFS_POWER cleared
[    2.766097] ACPI: AC Adapter [ACAD] (off-line)
[    2.826203] ACPI: Deprecated procfs I/F for AC is loaded, please retry with CONFIG_ACPI_PROCFS_POWER cleared
[    2.826231] ------------[ cut here ]------------
[    2.826253] WARNING: at fs/proc/generic.c:586 proc_register+0x10f/0x1f0()
[    2.826261] Hardware name: Inspiron 1090
[    2.826268] proc_dir_entry 'ac_adapter/ACAD' already registered
[    2.826275] Modules linked in:
[    2.826287] Pid: 1, comm: swapper Not tainted 3.1.6-1.fc16.x86_64 #1
[    2.826295] Call Trace:
[    2.826315]  [<ffffffff8106b73f>] warn_slowpath_common+0x7f/0xc0
[    2.826359]  [<ffffffff8106b836>] warn_slowpath_fmt+0x46/0x50
[    2.826373]  [<ffffffff811d5bbf>] proc_register+0x10f/0x1f0
[    2.826387]  [<ffffffff811d619f>] proc_mkdir_mode+0x3f/0x60
[    2.826400]  [<ffffffff811d61d6>] proc_mkdir+0x16/0x20
[    2.826414]  [<ffffffff8132b5e8>] acpi_ac_add+0xc2/0x1b8
[    2.826429]  [<ffffffff81308d22>] acpi_device_probe+0x4e/0x11c
[    2.826444]  [<ffffffff8138ba2c>] driver_probe_device+0x9c/0x2b0
[    2.826456]  [<ffffffff8138bceb>] __driver_attach+0xab/0xb0
[    2.826469]  [<ffffffff8138bc40>] ? driver_probe_device+0x2b0/0x2b0
[    2.826483]  [<ffffffff8138bc40>] ? driver_probe_device+0x2b0/0x2b0
[    2.826497]  [<ffffffff8138a8ac>] bus_for_each_dev+0x5c/0x90
[    2.826510]  [<ffffffff8138b63e>] driver_attach+0x1e/0x20
[    2.826523]  [<ffffffff8138b240>] bus_add_driver+0x1b0/0x2a0
[    2.826539]  [<ffffffff81bb06ad>] ? acpi_initialize_subsystem+0xa9/0xa9
[    2.826551]  [<ffffffff8138c246>] driver_register+0x76/0x140
[    2.826563]  [<ffffffff81063070>] ? try_to_wake_up+0x290/0x290
[    2.826576]  [<ffffffff81bb06ad>] ? acpi_initialize_subsystem+0xa9/0xa9
[    2.826590]  [<ffffffff813095a4>] acpi_bus_register_driver+0x43/0x45
[    2.826603]  [<ffffffff81bb06e0>] acpi_ac_init+0x33/0x4f
[    2.826617]  [<ffffffff81002042>] do_one_initcall+0x42/0x180
[    2.826632]  [<ffffffff81b7ecf3>] kernel_init+0xde/0x158
[    2.826644]  [<ffffffff810636d7>] ? schedule_tail+0x27/0xb0
[    2.826659]  [<ffffffff815dee34>] kernel_thread_helper+0x4/0x10
[    2.826672]  [<ffffffff81b7ec15>] ? start_kernel+0x3c0/0x3c0
[    2.826685]  [<ffffffff815dee30>] ? gs_change+0x13/0x13
[    2.826706] ---[ end trace 82bc55f036371117 ]---
[    2.826762] ------------[ cut here ]------------
[    2.826776] WARNING: at fs/sysfs/dir.c:455 sysfs_add_one+0xc1/0xf0()
[    2.826784] Hardware name: Inspiron 1090
[    2.826790] sysfs: cannot create duplicate filename '/class/power_supply/ACAD'
[    2.826797] Modules linked in:
[    2.826808] Pid: 1, comm: swapper Tainted: G        W   3.1.6-1.fc16.x86_64 #1
[    2.826815] Call Trace:
[    2.826829]  [<ffffffff8106b73f>] warn_slowpath_common+0x7f/0xc0
[    2.826842]  [<ffffffff8106b836>] warn_slowpath_fmt+0x46/0x50
[    2.826855]  [<ffffffff811e2411>] sysfs_add_one+0xc1/0xf0
[    2.826869]  [<ffffffff811e2dc5>] sysfs_do_create_link+0x135/0x220
[    2.826881]  [<ffffffff811e1561>] ? sysfs_add_file_mode+0x61/0xb0
[    2.826895]  [<ffffffff811e2ec3>] sysfs_create_link+0x13/0x20
[    2.826908]  [<ffffffff81388b06>] device_add+0x2b6/0x680
[    2.826923]  [<ffffffff812aa3e8>] ? kobject_set_name+0x38/0x40
[    2.826938]  [<ffffffff814514a8>] power_supply_register+0xf8/0x140
[    2.826952]  [<ffffffff8132b6ae>] acpi_ac_add+0x188/0x1b8
[    2.826965]  [<ffffffff81308d22>] acpi_device_probe+0x4e/0x11c
[    2.826979]  [<ffffffff8138ba2c>] driver_probe_device+0x9c/0x2b0
[    2.826991]  [<ffffffff8138bceb>] __driver_attach+0xab/0xb0
[    2.827004]  [<ffffffff8138bc40>] ? driver_probe_device+0x2b0/0x2b0
[    2.827052]  [<ffffffff8138bc40>] ? driver_probe_device+0x2b0/0x2b0
[    2.827066]  [<ffffffff8138a8ac>] bus_for_each_dev+0x5c/0x90
[    2.827079]  [<ffffffff8138b63e>] driver_attach+0x1e/0x20
[    2.827092]  [<ffffffff8138b240>] bus_add_driver+0x1b0/0x2a0
[    2.827106]  [<ffffffff81bb06ad>] ? acpi_initialize_subsystem+0xa9/0xa9
[    2.827118]  [<ffffffff8138c246>] driver_register+0x76/0x140
[    2.827129]  [<ffffffff81063070>] ? try_to_wake_up+0x290/0x290
[    2.827143]  [<ffffffff81bb06ad>] ? acpi_initialize_subsystem+0xa9/0xa9
[    2.827156]  [<ffffffff813095a4>] acpi_bus_register_driver+0x43/0x45
[    2.827169]  [<ffffffff81bb06e0>] acpi_ac_init+0x33/0x4f
[    2.827182]  [<ffffffff81002042>] do_one_initcall+0x42/0x180
[    2.827195]  [<ffffffff81b7ecf3>] kernel_init+0xde/0x158
[    2.827207]  [<ffffffff810636d7>] ? schedule_tail+0x27/0xb0
[    2.827221]  [<ffffffff815dee34>] kernel_thread_helper+0x4/0x10
[    2.827234]  [<ffffffff81b7ec15>] ? start_kernel+0x3c0/0x3c0
[    2.827247]  [<ffffffff815dee30>] ? gs_change+0x13/0x13
[    2.827255] ---[ end trace 82bc55f036371118 ]---
[    2.887119] ACPI: AC Adapter [ACAD] (off-line)

Re: 3.1 Panic on power event

[Bjorn Helgaas]

On Tue, Jan 10, 2012 at 9:59 AM, Josh Boyer <jwboyer@redhat.com> wrote:
> We've had a report [1] from a user that the 3.1.6 kernel panics if they
> plug or unplug their machine.  The oops seems to come from an issue
> somewhere in the acpi notification stack.  The relevant backtrace
> functions are (transcribed from a jpeg):
>
> acpi_ac_notify
> acpi_device_notify
> acpi_ev_notify_dispatch
> acpi_os_execute_deferred

Images 1 & 2 from the bugzilla also show kobject_uevent(), so my guess
is the problem is in the ACPI ac driver, e.g., maybe
ac->charger.dev->kobj isn't being initialized correctly, so we blow up
when acpi_ac_notify() calls kobject_uevent().  Added cc: for recent ac
changes.

Bjorn
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [3.0? -> 3.1.6 regression] Inspiron 1090: Panic on power event (Doubled AC adapter)

[Jonathan Nieder]

Hi ACPI folks,

Bjorn Helgaas wrote:
> On Tue, Jan 10, 2012 at 9:59 AM, Josh Boyer <jwboyer@redhat.com> wrote:

>> We've had a report [1] from a user that the 3.1.6 kernel panics if they
>> plug or unplug their machine.  The oops seems to come from an issue
>> somewhere in the acpi notification stack.  The relevant backtrace
>> functions are (transcribed from a jpeg):
>>
>> acpi_ac_notify
>> acpi_device_notify
>> acpi_ev_notify_dispatch
>> acpi_os_execute_deferred
>
> Images 1 & 2 from the bugzilla also show kobject_uevent(), so my guess
> is the problem is in the ACPI ac driver, e.g., maybe
> ac->charger.dev->kobj isn't being initialized correctly, so we blow up
> when acpi_ac_notify() calls kobject_uevent().  Added cc: for recent ac
> changes.
[...]
>> [1] https://bugzilla.redhat.com/show_bug.cgi?id=772730

More reports:

 https://bugzilla.kernel.org/show_bug.cgi?id=38782 (includes DSDT fix)
 http://bugs.debian.org/661995
 https://bugs.launchpad.net/ubuntu/+source/linux/+bug/414541

Dell doesn't take emailed reports re the Inspiron 1090 BIOS so
Giovanni Biscuolo (cc-ed) reported it to

 http://en.community.dell.com/support-forums/laptop/f/3518/t/19439525.aspx

I wonder:

 - could Linux diagnose the duplicate ACAD in a more helpful way?
 - is it possible to work around, for example by letting the last AC
   adaptor win on the Inspiron 1090?
 - what does Windows do?

Hope that helps,
Jonathan
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Inspiron 1090: Panic on power event (Doubled AC adapter)

[Jonathan Nieder]

Jonathan Nieder wrote:

> [Subject: Re: [3.0? -> 3.1.6 regression] Inspiron 1090: Panic on
> power event (Doubled AC adapter)]

Hmm, it's older than that.  Might not be a regression.  Sorry for the
confusion.

Jonathan