Kernel bug mm/page_alloc.c

Kernel bug mm/page_alloc.c

[Michael Gilroy]

Hi,
       I get a repeatable error due to failures in mm/page_alloc.c, errors 
occur during heavy utilisation of the file-system with bonnie++. I'm using 
an experimental driver for the RAID 6 operations however as the error 
message below indicates that the bug is being caused elsewhere I wonder if 
this is a bug in the kernel. If not can anyone suggest a probable cause or 
method to help locate the cause of this bug?
kernel BUG at mm/page_alloc.c:761!

invalid operand: 0000 [#1]

Modules linked in: raid6 xor radeon drm parport_pc lp parport autofs4 
i2c_dev i2 c_core sunrpc ipt_REJECT ipt_state ip_conntrack iptable_filter 
ip_tables dm_mod video button battery ac ipv6 ohci_hcd hw_random sata_sil24 
tg3 floppy ext3 jbd s ata_sil libata sd_mod scsi_mod

CPU: 0

EIP: 0060:[<c0136172>] Not tainted VLI

EFLAGS: 00010202 (2.6.15.2)

EIP is at buffered_rmqueue+0xbd/0x1a3

eax: 00000001 ebx: c0319720 ecx: 00001000 edx: 00000001

esi: c0319720 edi: c0319740 ebp: eb9cfd24 esp: eb9cfcfc

ds: 007b es: 007b ss: 0068

Process bonnie++ (pid: 4548, threadinfo=eb9cf000 task=ecd99030)

Stack: 00001000 00000001 00000000 c1210000 00000246 000200d2 00000000 
c03199ec

00000044 00000000 eb9cfd44 c0136367 00000003 00000000 000200d2 c03199e8

000200d2 eb9cfdb0 eb9cfd74 c01363d8 00000044 00000000 ecd99030 00000010

Call Trace:

[<c0103078>] show_stack+0x76/0x7e

[<c010317f>] show_registers+0xe8/0x150

[<c010331a>] die+0xbe/0x12f

[<c02c5122>] do_trap+0x7c/0x96

[<c010357b>] do_invalid_op+0x95/0x9c

[<c0102d7f>] error_code+0x4f/0x54

[<c0136367>] get_page_from_freelist+0x74/0x8c

[<c01363d8>] __alloc_pages+0x59/0x271

[<c01341cd>] generic_file_buffered_write+0x157/0x4b9

[<c01348d5>] __generic_file_aio_write_nolock+0x3a6/0x3d1

[<c01349f0>] __generic_file_write_nolock+0x74/0x8a

[<c0134b8f>] generic_file_write+0x49/0xaf

[<c014cd73>] vfs_write+0xa9/0x14e

[<c014ceb3>] sys_write+0x3b/0x60

[<c0102afb>] sysenter_past_esp+0x54/0x75

Code: f0 e8 e3 fb ff ff 89 45 e4 8b 55 e8 89 d8 e8 26 ee 18 00 83 7d e4 00 
0f 84 e9 00 00 00 8b 55 e4 89 f0 e8 5c f7 ff ff 85 c0 74 08 <0f> 0b f9 02 f5 
88 2d c0 8b 96 38 01 00 00 bf 28 00 00 00 8d 82




Please CC me on any replies as I'm not subscribed to this list.

thanks
_______________________
Michael Gilroy
Research Engineer
A2E Limited
Adaptive House
Quarrywood Court
Livingston
EH54 6AX
Scotland 

Re: Kernel bug mm/page_alloc.c

[Ram Gupta]

On 2/15/06, Michael Gilroy <mgilroy@a2etech.com> wrote:
> Hi,
>        I get a repeatable error due to failures in mm/page_alloc.c, errors
> occur during heavy utilisation of the file-system with bonnie++. I'm using
> an experimental driver for the RAID 6 operations however as the error

This problem seems related to intel F00F BUG. Is CONFIG_X86_F00F_BUG
enabled in your current kernel. If so then can you try with this
option disabled?

Regards
Ram Gupta

Re: Kernel bug mm/page_alloc.c

[Michael Gilroy]

I'm using an AMD opteron so I can't see how the F00F bug could be affecting 
me. I don't have that option in my .config file either. Any other 
suggestions?

thanks,

mike 

Re: Kernel bug mm/page_alloc.c

[Ram Gupta]

On 2/16/06, Michael Gilroy <mgilroy@a2etech.com> wrote:
> I'm using an AMD opteron so I can't see how the F00F bug could be affecting
> me. I don't have that option in my .config file either. Any other
> suggestions?
>
> thanks,
>
> mike

Does this problem pop up without the RAID6 driver also. That may be 
helpful in isolating the bug.

regards
Ram Gupta
>
>

Re: Kernel bug mm/page_alloc.c

[Michael Gilroy]

Hi,
     apologies for the delay in replying I've been out of the country.
I'm using the 2.6.15.2 kernel. I know that the bug is only appearing when
I'm using the raid6 driver in combination with my hardware. I've yet to
narrow it down any further though and would appreciate it if anyone could
give me a better idea as to how to narrow down the cause of the problem.

regards,

mike

----- Original Message ----- 
From: "Ram Gupta" <ram.gupta5@gmail.com>
To: "Michael Gilroy" <mgilroy@a2etech.com>
Cc: <linux-kernel@vger.kernel.org>
Sent: Thursday, February 16, 2006 9:30 PM
Subject: Re: Kernel bug mm/page_alloc.c


On 2/16/06, Michael Gilroy <mgilroy@a2etech.com> wrote:
> I'm using an AMD opteron so I can't see how the F00F bug could be 
> affecting
> me. I don't have that option in my .config file either. Any other
> suggestions?
>
> thanks,
>
> mike

Does this problem pop up without the RAID6 driver also. That may be
helpful in isolating the bug.

regards
Ram Gupta
>
>