Prev: Irish 2010 Winner
Next: [PATCH 0/4] Various intel small device drivers
From: Christian Hartmann on 14 Apr 2010 09:20
Hi,

I have the same problem on my thinkpad T30 since this week:

The kernel boots very slowly until the message:

* Loading kernel modules...

there my kernel hangs there forever now.

The first problem is shown after the start of the udev (:

* Waiting for /dev to be fully populated...

[ 18.273664] BUG: unable to handle kernel paging request at 0033393a
[ 18.274036] IP: [<c027d8f5>] acpi_ex_read_data_from_field+0x117/0x14e
[ 18.274249] *pde = 00000000
[ 18.274446] Oops: 0000 [#1]
[ 18.274625] last sysfs file:
/sys/devices/pci0000:00/0000:00:1f.1/ide0/0.0/block/hda/dev
[ 18.274625] Modules linked in : battery ac processor rng_core
button uhci_hcd ehci_hcd thermal usbcore [last unloaded:
scsi_wait_scan]
[ 18.274625]
[ 18.274625] Pid: 1077, comm: async/0 Not tained 2.6.34-rc4 #61
2366DG3/2366DG3
[ 18.274625] EIP: 0060:[<c027d8f5>] EFLAGS: 00010246 CPU:0
[ 18.274625] EIP is at acpi_ex_read_data_from_field+0x117/0x14e
[ 18.274625] EAX: 00000000 EBX: 00333931 ECX: 00000000 EDX: 00000000
[ 18.274625] ESI: 00000000 EDI: efa4dd70 EBP: efa4dd64 ESP: efa4dd50
[ 18.274625] DS: 007b ES: 007b FS: 0000 GS: 00e0 SS: 0068
[ 18.274625] Process async/0 (pid:1077, ti=efa4c000 task= f70a2850
task.ti=efa4c000
[ 18.274625] Stack: <me do not want to type it all...>
[ 18.274625] Call Trace:
[ 18.274625] [<c02823be>] ? acpi_ex_resolve_node_to_value+0x166/0x1dc
[ 18.274625] [<c027e226>] ? acpi_ex_resolve_to_value+0x65/0x6f
[ 18.274625] [<c0279065>] ? acpi_ds_evaluate_name_path+0x71/0xdc
.....<me do not want to type it all...>...
[ 18.274625] [<c0102e3e>] ? kernel_thread_helper+0x6/0x10
[ 18.274625] Code: c6 74 4e 0f b6 0d 7d a6 7e c0 8d 50 0c 0f b6 43
09 89 55 f0 89 4d ec e8 93 42 00 00 8b 55 f0 89 d8 8b 4d ec e8 9e 12
00 00 89 c2 <0f> b6 43 09 89 55 f0 e8 44 42 00 00 8b 55 f0 85 d2 74 0f
89 f0
[ 18.274625] EIP: [<c027d8f5>] ?
acpi_ex_read_data_from_field+0x117/0x14e SS: ESP 0068:efa4dd50
[ 18.274625] CR2: 000000000033393a
[ 18.274625] ---[ end trace da97996273142d42 ]---
[ 18.274625] bus: 'pcmcia': registered
....
* Starting early crypto disks...
* Cleaning up ifdown...
* Loading kernel modules...
--- FREEEZE ---

I can boot into the lap with the stable 2.6.32.11 and can give other
infos like the current kernelconfig of this laptop. But I will wait
for the reboot, leaving the message still open, if anybody needs more
info.


sincerelly your
Chris

2010/4/14 Sergey Senozhatsky <sergey.senozhatsky(a)gmail.com>:
>
> Hello,
> Kernel oops on early stages. The below trace is "copy-paste" from the paper
> since syslog is not runing yet.
>
> kernel: .34-rc4-git1 /* also tested on .34-rc4 .34-rc3-git10 */
> laptop: ASUS f3jc x86
>
>
> ds: 007b es:007b fs:00d8 gs:00e8 ss:0068
> process: async/0
> stack:
> � � � �0000000 c144c708 f66cdd20 c109fc0b f723c1b8 f71fcec0 f71aa208 00000020
> � � � �<0> � � c11c4d40 00001000 000000b6 c12f61e0 c12f61bc f723c1b8 f71aa208
> � � � �f71fcec0
> � � � �<0> � � f66cdd54 c11c4d58 00000020 f66cdd64 f73a0f84 f723c1b8 f70cdc70
> � � � �f66cdd74
>
> call trace:
> � � � �__kmalloc+0xe1
> � � � �acpi_ex_read_data_from_field+0x156
> � � � �acpi_ex_read_data_from_field+0x16e
> � � � �acpi_ex_resolve_node_to_value
> � � � �acpi_ex_resolve_to_value
> � � � �acpi_ds_evaluate_name_path
> � � � �acpi_ds_exec_end_op
> � � � �acpi_ps_parse_loop
> � � � �trace_hardirqs_on
> � � � �acpi_ps_parse_aml
> � � � �acpi_ps_execute_method
> � � � �acpi_ns_eveluate
> � � � �acpi_evaluate_object
> � � � �acpi_battery_update
> � � � �acpi_battery_add
> � � � �acpi_device_probe
> � � � �driver_probe_device
> � � � �__driver_attach
> � � � �bus_for_each_dev
> � � � �driver_attach
> � � � �__driver_attach
> � � � �bus_add_driver
> � � � �driver_register
> � � � �trace_hardirqs_on_caller
> � � � �acpi_bus_register_driver
> � � � �acpi_battery_init_async
> � � � �async_thread
> � � � �default_wake_function
> � � � �async_thread
> � � � �kthread
> � � � �kthread
> � � � �kernel_thread_helper
>
> code:
> � � � �fb 00 00 b8 0b 00 00 00 83 c4 14 e9 7a 01 � 00 00 31 c0 8b 7d d4 8b 4d dc 31 d2 f3 aa 6a 00
> � � � �0f b6 4b 1d 8b 43 10 8d 74 01 ff 89 f0 <f7> f1 31 d2 89 45 cc 0f b6 43 1c 8d 04 06 8d 75 ec
> � � � �f7 f1 31 d2
>
> eip: <c11c66dd> acpi_ex_extract_from_field+0x5c/0x1c3 ss:esp 0068:f66cdcf8
>
>
> � � � �Sergey



--
--
.......oooO..............
......(....).....Oooo...
.......)../.....(....)....
......(_/.......)../.....
................(_/.......
.... I WAS ..............
........... HERE...... ;)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Luca Tettamanti on 14 Apr 2010 09:20
Hello,
I'm seeing a similar bug on my notebook (Asus F3Sa) with current git.
-rc4 does not work, while -rc3 does work.

The error (at least in my case) is a division by zero in
acpi_ex_insert_into_field.
Disassembly near the faulty instruction (div %esi) looks like this:

0x000000000001cbee <+205>: 74 09 je 0x1cbf9 <acpi_ex_insert_into_field+216>
0x000000000001cbf0 <+207>: 40 88 f1 mov %sil,%cl
0x000000000001cbf3 <+210>: 49 d3 e5 shl %cl,%r13
0x000000000001cbf6 <+213>: 49 f7 d5 not %r13
0x000000000001cbf9 <+216>: 0f b6 4b 24 movzbl 0x24(%rbx),%ecx
0x000000000001cbfd <+220>: 40 0f b6 f6 movzbl %sil,%esi
0x000000000001cc01 <+224>: 49 83 c9 ff or $0xffffffffffffffff,%r9
0x000000000001cc05 <+228>: 8b 7b 18 mov 0x18(%rbx),%edi
0x000000000001cc08 <+231>: 49 d3 e1 shl %cl,%r9
0x000000000001cc0b <+234>: 8d 7c 3e ff lea -0x1(%rsi,%rdi,1),%edi
0x000000000001cc0f <+238>: 31 d2 xor %edx,%edx
0x000000000001cc11 <+240>: 89 f8 mov %edi,%eax
0x000000000001cc13 <+242>: 4d 21 e9 and %r13,%r9
0x000000000001cc16 <+245>: f7 f6 div %esi
0x000000000001cc18 <+247>: 41 bb 01 00 00 00 mov $0x1,%r11d
0x000000000001cc1e <+253>: 89 45 b4 mov %eax,-0x4c(%rbp)
0x000000000001cc21 <+256>: 31 d2 xor %edx,%edx
0x000000000001cc23 <+258>: 8d 04 0f lea (%rdi,%rcx,1),%eax
0x000000000001cc26 <+261>: 0f b6 4b 0f movzbl 0xf(%rbx),%ecx

Register dump confirms that ESI is 0. The div seems to be generated for
ACPI_ROUND_UP_TO:

865 datum_count = ACPI_ROUND_UP_TO(obj_desc->common_field.bit_length,
0x0000000000000d11 <+220>: movzbl %sil,%esi
0x0000000000000d19 <+228>: mov 0x18(%rbx),%edi
0x0000000000000d1f <+234>: lea -0x1(%rsi,%rdi,1),%edi
0x0000000000000d23 <+238>: xor %edx,%edx
0x0000000000000d25 <+240>: mov %edi,%eax
0x0000000000000d2a <+245>: div %esi
0x0000000000000d32 <+253>: mov %eax,-0x4c(%rbp)

(recompiled with DEBUG_INFO)

Stack trace:
acpi_ex_insert_into_field
acpi_ex_write_data_to_field
acpi_ex_store_object_to_node
acpi_ex_store
acpi_ex_opcode_1A_1T_1R
acpi_ds_exec_end_op
acpi_ps_parse_loop
acpi_ps_parse_aml
acpi_ps_execute_method
acpi_ns_evaluate
acpi_evaluate_object
acpi_battery_update [battery]
acpi_battery_add [battery]
acpi_device_probe
driver_probe_device
bus_for_each_dev
driver_attach
bus_add_driver
driver_register
acpi_bus_register_driver
acpi_battery_init_async [battery]


Compiler is:
$ gcc --version
gcc (Debian 4.4.3-7) 4.4.3

Luca
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Tero Roponen on 14 Apr 2010 10:00

On Wed, 14 Apr 2010, Luca Tettamanti wrote:

> Hello,
> I'm seeing a similar bug on my notebook (Asus F3Sa) with current git.
> -rc4 does not work, while -rc3 does work.
>
> The error (at least in my case) is a division by zero in
> acpi_ex_insert_into_field.

I have a similar problem with Asus M50Vm. git-bisect found
the following commit and reverting it does indeed fix the
problem.

commit dadf28a10c3eb29421837a2e413ab869ebd9e168
Author: Alexey Starikovskiy <astarikovskiy(a)suse.de>
Date: Wed Mar 17 13:14:13 2010 -0400

ACPI: EC: Allow multibyte access to EC

http://bugzilla.kernel.org/show_bug.cgi?id=14667

Signed-off-by: Alexey Starikovskiy <astarikovskiy(a)suse.de>
Signed-off-by: Len Brown <len.brown(a)intel.com>

--
Tero Roponen
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
From: Tero Roponen on 14 Apr 2010 10:10

On Wed, 14 Apr 2010, Alexey Starikovskiy wrote:

> 14.04.2010 17:50, Tero Roponen пишет:
> >
> > On Wed, 14 Apr 2010, Luca Tettamanti wrote:
> >
> > > Hello,
> > > I'm seeing a similar bug on my notebook (Asus F3Sa) with current git.
> > > -rc4 does not work, while -rc3 does work.
> > >
> > > The error (at least in my case) is a division by zero in
> > > acpi_ex_insert_into_field.
> >
> > I have a similar problem with Asus M50Vm. git-bisect found
> > the following commit and reverting it does indeed fix the
> > problem.
> >
> > commit dadf28a10c3eb29421837a2e413ab869ebd9e168
> > Author: Alexey Starikovskiy<astarikovskiy(a)suse.de>
> > Date: Wed Mar 17 13:14:13 2010 -0400
> >
> > ACPI: EC: Allow multibyte access to EC
> >
> > http://bugzilla.kernel.org/show_bug.cgi?id=14667
> >
> > Signed-off-by: Alexey Starikovskiy<astarikovskiy(a)suse.de>
> > Signed-off-by: Len Brown<len.brown(a)intel.com>
> >
> Please see the patch in https://bugzilla.kernel.org/show_bug.cgi?id=15749

Work for me. Thanks.

Tested-by: Tero Roponen <tero.roponen(a)gmail.com>
From: Alexey Starikovskiy on 14 Apr 2010 10:20
14.04.2010 17:50, Tero Roponen пишет:
>
> On Wed, 14 Apr 2010, Luca Tettamanti wrote:
>
>> Hello,
>> I'm seeing a similar bug on my notebook (Asus F3Sa) with current git.
>> -rc4 does not work, while -rc3 does work.
>>
>> The error (at least in my case) is a division by zero in
>> acpi_ex_insert_into_field.
>
> I have a similar problem with Asus M50Vm. git-bisect found
> the following commit and reverting it does indeed fix the
> problem.
>
> commit dadf28a10c3eb29421837a2e413ab869ebd9e168
> Author: Alexey Starikovskiy<astarikovskiy(a)suse.de>
> Date: Wed Mar 17 13:14:13 2010 -0400
>
> ACPI: EC: Allow multibyte access to EC
>
> http://bugzilla.kernel.org/show_bug.cgi?id=14667
>
> Signed-off-by: Alexey Starikovskiy<astarikovskiy(a)suse.de>
> Signed-off-by: Len Brown<len.brown(a)intel.com>
>
Please see the patch in https://bugzilla.kernel.org/show_bug.cgi?id=15749

Thanks,
Alex.


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo(a)vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
 |  Next  |  Last
Pages: 1 2
Prev: Irish 2010 Winner
Next: [PATCH 0/4] Various intel small device drivers