strange stuff in dmesg

strange stuff in dmesg

[BuraphaLinux Server]

[-- Attachment #1: Type: text/plain, Size: 1504 bytes --]

On my Dell OptiPlex 330 machines with kernel 2.6.32.2 I get a strange
WARNING.  Do I need to worry?  Here is the warning:

[    0.000000] ------------[ cut here ]------------
[    0.000000] WARNING: at mm/page_alloc.c:1805
__alloc_pages_nodemask+0x1b6/0x730()
[    0.000000] Hardware name: OptiPlex 330
[    0.000000] Modules linked in:
[    0.000000] Pid: 0, comm: swapper Not tainted 2.6.32.2 #1
[    0.000000] Call Trace:
[    0.000000]  [<ffffffff8108e806>] ? __alloc_pages_nodemask+0x1b6/0x730
[    0.000000]  [<ffffffff81043f68>] warn_slowpath_common+0x78/0xd0
[    0.000000]  [<ffffffff81043fcf>] warn_slowpath_null+0xf/0x20
[    0.000000]  [<ffffffff8108e806>] __alloc_pages_nodemask+0x1b6/0x730
[    0.000000]  [<ffffffff810b4e28>] alloc_pages_current+0x78/0xf0
[    0.000000]  [<ffffffff8108da69>] __get_free_pages+0x9/0x50
[    0.000000]  [<ffffffff810bb912>] __kmalloc+0x112/0x120
[    0.000000]  [<ffffffff8146e773>] vgacon_scrollback_startup+0x13/0x70
[    0.000000]  [<ffffffff811d29b3>] vgacon_startup+0x2a3/0x420
[    0.000000]  [<ffffffff816fc556>] con_init+0x1b/0x230
[    0.000000]  [<ffffffff816fba00>] console_init+0x22/0x42
[    0.000000]  [<ffffffff816d4b8f>] start_kernel+0x240/0x3be
[    0.000000]  [<ffffffff816d4289>] x86_64_start_reservations+0x99/0xb9
[    0.000000]  [<ffffffff816d4389>] x86_64_start_kernel+0xe0/0xf2
[    0.000000] ---[ end trace 4eaa2a86a8e2da22 ]---

I also got it on 2.6.31.9, but had waited to ask hoping 2.6.32.2 would fix it.

Attached is my config

[-- Attachment #2: config.gz --]
[-- Type: application/x-gzip, Size: 18239 bytes --]

Re: strange stuff in dmesg

[BuraphaLinux Server]

On 12/24/09, Justin P. Mattock <justinmattock@gmail.com> wrote:
> On 12/24/09 02:03, BuraphaLinux Server wrote:
>> On my Dell OptiPlex 330 machines with kernel 2.6.32.2 I get a strange
>> WARNING.  Do I need to worry?  Here is the warning:
>>
>> [    0.000000] ------------[ cut here ]------------
>> [    0.000000] WARNING: at mm/page_alloc.c:1805
>> __alloc_pages_nodemask+0x1b6/0x730()
>> [    0.000000] Hardware name: OptiPlex 330
>> [    0.000000] Modules linked in:
>> [    0.000000] Pid: 0, comm: swapper Not tainted 2.6.32.2 #1
>> [    0.000000] Call Trace:
>> [    0.000000]  [<ffffffff8108e806>] ? __alloc_pages_nodemask+0x1b6/0x730
>> [    0.000000]  [<ffffffff81043f68>] warn_slowpath_common+0x78/0xd0
>> [    0.000000]  [<ffffffff81043fcf>] warn_slowpath_null+0xf/0x20
>> [    0.000000]  [<ffffffff8108e806>] __alloc_pages_nodemask+0x1b6/0x730
>> [    0.000000]  [<ffffffff810b4e28>] alloc_pages_current+0x78/0xf0
>> [    0.000000]  [<ffffffff8108da69>] __get_free_pages+0x9/0x50
>> [    0.000000]  [<ffffffff810bb912>] __kmalloc+0x112/0x120
>> [    0.000000]  [<ffffffff8146e773>] vgacon_scrollback_startup+0x13/0x70
>> [    0.000000]  [<ffffffff811d29b3>] vgacon_startup+0x2a3/0x420
>> [    0.000000]  [<ffffffff816fc556>] con_init+0x1b/0x230
>> [    0.000000]  [<ffffffff816fba00>] console_init+0x22/0x42
>> [    0.000000]  [<ffffffff816d4b8f>] start_kernel+0x240/0x3be
>> [    0.000000]  [<ffffffff816d4289>] x86_64_start_reservations+0x99/0xb9
>> [    0.000000]  [<ffffffff816d4389>] x86_64_start_kernel+0xe0/0xf2
>> [    0.000000] ---[ end trace 4eaa2a86a8e2da22 ]---
>>
>> I also got it on 2.6.31.9, but had waited to ask hoping 2.6.32.2 would fix
>> it.
>>
>> Attached is my config
>>
> was there a kernel that did not do this?
> if so can you try a bisect on this?
>
> Justin P. Mattock
>

It took a while, but I have verified that 2.6.30.10 works without any
message, and 2.6.31 has the error message (but otherwise seems to run
ok).  The hex codes are different, but the function names match and
are in the same places.

Does it have to be git bisect, or will trying the 2.6.31rc[1-9] be
enough instead?  The Documentation/BUG-HUNTING does not give detailed
enough instructions for me to do the bisect thing.

Re: strange stuff in dmesg

[Randy Dunlap]

On Thu, 24 Dec 2009 22:22:41 +0700 BuraphaLinux Server wrote:

> On 12/24/09, Justin P. Mattock <justinmattock@gmail.com> wrote:
> > On 12/24/09 02:03, BuraphaLinux Server wrote:
> >> On my Dell OptiPlex 330 machines with kernel 2.6.32.2 I get a strange
> >> WARNING.  Do I need to worry?  Here is the warning:
> >>
> >> [    0.000000] ------------[ cut here ]------------
> >> [    0.000000] WARNING: at mm/page_alloc.c:1805
> >> __alloc_pages_nodemask+0x1b6/0x730()
> >> [    0.000000] Hardware name: OptiPlex 330
> >> [    0.000000] Modules linked in:
> >> [    0.000000] Pid: 0, comm: swapper Not tainted 2.6.32.2 #1
> >> [    0.000000] Call Trace:
> >> [    0.000000]  [<ffffffff8108e806>] ? __alloc_pages_nodemask+0x1b6/0x730
> >> [    0.000000]  [<ffffffff81043f68>] warn_slowpath_common+0x78/0xd0
> >> [    0.000000]  [<ffffffff81043fcf>] warn_slowpath_null+0xf/0x20
> >> [    0.000000]  [<ffffffff8108e806>] __alloc_pages_nodemask+0x1b6/0x730
> >> [    0.000000]  [<ffffffff810b4e28>] alloc_pages_current+0x78/0xf0
> >> [    0.000000]  [<ffffffff8108da69>] __get_free_pages+0x9/0x50
> >> [    0.000000]  [<ffffffff810bb912>] __kmalloc+0x112/0x120
> >> [    0.000000]  [<ffffffff8146e773>] vgacon_scrollback_startup+0x13/0x70
> >> [    0.000000]  [<ffffffff811d29b3>] vgacon_startup+0x2a3/0x420
> >> [    0.000000]  [<ffffffff816fc556>] con_init+0x1b/0x230
> >> [    0.000000]  [<ffffffff816fba00>] console_init+0x22/0x42
> >> [    0.000000]  [<ffffffff816d4b8f>] start_kernel+0x240/0x3be
> >> [    0.000000]  [<ffffffff816d4289>] x86_64_start_reservations+0x99/0xb9
> >> [    0.000000]  [<ffffffff816d4389>] x86_64_start_kernel+0xe0/0xf2
> >> [    0.000000] ---[ end trace 4eaa2a86a8e2da22 ]---
> >>
> >> I also got it on 2.6.31.9, but had waited to ask hoping 2.6.32.2 would fix
> >> it.
> >>
> >> Attached is my config
> >>
> > was there a kernel that did not do this?
> > if so can you try a bisect on this?
> >
> > Justin P. Mattock
> >
> 
> It took a while, but I have verified that 2.6.30.10 works without any
> message, and 2.6.31 has the error message (but otherwise seems to run
> ok).  The hex codes are different, but the function names match and
> are in the same places.
> 
> Does it have to be git bisect, or will trying the 2.6.31rc[1-9] be
> enough instead?  The Documentation/BUG-HUNTING does not give detailed
> enough instructions for me to do the bisect thing.

"git bisect" can point directly at the kernel patch that causes the
problem (well, most of the time it can do that), whereas just saying
2.6.31-rc[1-9] will just tell us that it was one of a few thousand
patches.  Not nearly as helpful.

Maybe this can help you:
http://www.kernel.org/pub/software/scm/git-core/docs/git-bisect-lk2009.html


---
~Randy

Re: strange stuff in dmesg

[Justin P. Mattock]

On 12/24/09 09:34, Randy Dunlap wrote:
> On Thu, 24 Dec 2009 22:22:41 +0700 BuraphaLinux Server wrote:
>
>> On 12/24/09, Justin P. Mattock<justinmattock@gmail.com>  wrote:
>>> On 12/24/09 02:03, BuraphaLinux Server wrote:
>>>> On my Dell OptiPlex 330 machines with kernel 2.6.32.2 I get a strange
>>>> WARNING.  Do I need to worry?  Here is the warning:
>>>>
>>>> [    0.000000] ------------[ cut here ]------------
>>>> [    0.000000] WARNING: at mm/page_alloc.c:1805
>>>> __alloc_pages_nodemask+0x1b6/0x730()
>>>> [    0.000000] Hardware name: OptiPlex 330
>>>> [    0.000000] Modules linked in:
>>>> [    0.000000] Pid: 0, comm: swapper Not tainted 2.6.32.2 #1
>>>> [    0.000000] Call Trace:
>>>> [    0.000000]  [<ffffffff8108e806>] ? __alloc_pages_nodemask+0x1b6/0x730
>>>> [    0.000000]  [<ffffffff81043f68>] warn_slowpath_common+0x78/0xd0
>>>> [    0.000000]  [<ffffffff81043fcf>] warn_slowpath_null+0xf/0x20
>>>> [    0.000000]  [<ffffffff8108e806>] __alloc_pages_nodemask+0x1b6/0x730
>>>> [    0.000000]  [<ffffffff810b4e28>] alloc_pages_current+0x78/0xf0
>>>> [    0.000000]  [<ffffffff8108da69>] __get_free_pages+0x9/0x50
>>>> [    0.000000]  [<ffffffff810bb912>] __kmalloc+0x112/0x120
>>>> [    0.000000]  [<ffffffff8146e773>] vgacon_scrollback_startup+0x13/0x70
>>>> [    0.000000]  [<ffffffff811d29b3>] vgacon_startup+0x2a3/0x420
>>>> [    0.000000]  [<ffffffff816fc556>] con_init+0x1b/0x230
>>>> [    0.000000]  [<ffffffff816fba00>] console_init+0x22/0x42
>>>> [    0.000000]  [<ffffffff816d4b8f>] start_kernel+0x240/0x3be
>>>> [    0.000000]  [<ffffffff816d4289>] x86_64_start_reservations+0x99/0xb9
>>>> [    0.000000]  [<ffffffff816d4389>] x86_64_start_kernel+0xe0/0xf2
>>>> [    0.000000] ---[ end trace 4eaa2a86a8e2da22 ]---
>>>>
>>>> I also got it on 2.6.31.9, but had waited to ask hoping 2.6.32.2 would fix
>>>> it.
>>>>
>>>> Attached is my config
>>>>
>>> was there a kernel that did not do this?
>>> if so can you try a bisect on this?
>>>
>>> Justin P. Mattock
>>>
>>
>> It took a while, but I have verified that 2.6.30.10 works without any
>> message, and 2.6.31 has the error message (but otherwise seems to run
>> ok).  The hex codes are different, but the function names match and
>> are in the same places.
>>
>> Does it have to be git bisect, or will trying the 2.6.31rc[1-9] be
>> enough instead?  The Documentation/BUG-HUNTING does not give detailed
>> enough instructions for me to do the bisect thing.
>
> "git bisect" can point directly at the kernel patch that causes the
> problem (well, most of the time it can do that), whereas just saying
> 2.6.31-rc[1-9] will just tell us that it was one of a few thousand
> patches.  Not nearly as helpful.
>
> Maybe this can help you:
> http://www.kernel.org/pub/software/scm/git-core/docs/git-bisect-lk2009.html
>
>
> ---
> ~Randy
>


yeah a bisect hopefully
points to the right location

if you need info I can try and assist you as much as
possible, and well as anybody else.

Justin P. Mattock

Re: strange stuff in dmesg

[Américo Wang]

On Thu, Dec 24, 2009 at 11:22 PM, BuraphaLinux Server
<buraphalinuxserver@gmail.com> wrote:
> On 12/24/09, Justin P. Mattock <justinmattock@gmail.com> wrote:
>> On 12/24/09 02:03, BuraphaLinux Server wrote:
>>> On my Dell OptiPlex 330 machines with kernel 2.6.32.2 I get a strange
>>> WARNING.  Do I need to worry?  Here is the warning:
>>>
>>> [    0.000000] ------------[ cut here ]------------
>>> [    0.000000] WARNING: at mm/page_alloc.c:1805
>>> __alloc_pages_nodemask+0x1b6/0x730()
>>> [    0.000000] Hardware name: OptiPlex 330
>>> [    0.000000] Modules linked in:
>>> [    0.000000] Pid: 0, comm: swapper Not tainted 2.6.32.2 #1
>>> [    0.000000] Call Trace:
>>> [    0.000000]  [<ffffffff8108e806>] ? __alloc_pages_nodemask+0x1b6/0x730
>>> [    0.000000]  [<ffffffff81043f68>] warn_slowpath_common+0x78/0xd0
>>> [    0.000000]  [<ffffffff81043fcf>] warn_slowpath_null+0xf/0x20
>>> [    0.000000]  [<ffffffff8108e806>] __alloc_pages_nodemask+0x1b6/0x730
>>> [    0.000000]  [<ffffffff810b4e28>] alloc_pages_current+0x78/0xf0
>>> [    0.000000]  [<ffffffff8108da69>] __get_free_pages+0x9/0x50
>>> [    0.000000]  [<ffffffff810bb912>] __kmalloc+0x112/0x120
>>> [    0.000000]  [<ffffffff8146e773>] vgacon_scrollback_startup+0x13/0x70
>>> [    0.000000]  [<ffffffff811d29b3>] vgacon_startup+0x2a3/0x420
>>> [    0.000000]  [<ffffffff816fc556>] con_init+0x1b/0x230
>>> [    0.000000]  [<ffffffff816fba00>] console_init+0x22/0x42
>>> [    0.000000]  [<ffffffff816d4b8f>] start_kernel+0x240/0x3be
>>> [    0.000000]  [<ffffffff816d4289>] x86_64_start_reservations+0x99/0xb9
>>> [    0.000000]  [<ffffffff816d4389>] x86_64_start_kernel+0xe0/0xf2
>>> [    0.000000] ---[ end trace 4eaa2a86a8e2da22 ]---
>>>
>>> I also got it on 2.6.31.9, but had waited to ask hoping 2.6.32.2 would fix
>>> it.
>>>
>>> Attached is my config
>>>
>> was there a kernel that did not do this?
>> if so can you try a bisect on this?
>>
>> Justin P. Mattock
>>
>
> It took a while, but I have verified that 2.6.30.10 works without any
> message, and 2.6.31 has the error message (but otherwise seems to run
> ok).  The hex codes are different, but the function names match and
> are in the same places.
>
> Does it have to be git bisect, or will trying the 2.6.31rc[1-9] be
> enough instead?  The Documentation/BUG-HUNTING does not give detailed
> enough instructions for me to do the bisect thing.

Hey, the problem is that you configured VGACON_SOFT_SCROLLBACK_SIZE
to be CONFIG_VGACON_SOFT_SCROLLBACK_SIZE=65536, which will be
65536Kb, i.e. 16384 pages, how huge! :)

We shoud limit the number of VGACON_SOFT_SCROLLBACK_SIZE. I will cook
a patch now...

Re: strange stuff in dmesg

[Dan Carpenter]

On Thu, Dec 24, 2009 at 05:03:58PM +0700, BuraphaLinux Server wrote:
> On my Dell OptiPlex 330 machines with kernel 2.6.32.2 I get a strange
> WARNING.  Do I need to worry?  Here is the warning:
> 
> [    0.000000] ------------[ cut here ]------------
> [    0.000000] WARNING: at mm/page_alloc.c:1805
> __alloc_pages_nodemask+0x1b6/0x730()

Memory allocations are failing because your scrollback
size is set to high (65 Mb).

In your config you have:
CONFIG_VGACON_SOFT_SCROLLBACK_SIZE=65536

The default in 2.6.32 is 64.

regards,
dan carpenter

> [    0.000000] Hardware name: OptiPlex 330
> [    0.000000] Modules linked in:
> [    0.000000] Pid: 0, comm: swapper Not tainted 2.6.32.2 #1
> [    0.000000] Call Trace:
> [    0.000000]  [<ffffffff8108e806>] ? __alloc_pages_nodemask+0x1b6/0x730
> [    0.000000]  [<ffffffff81043f68>] warn_slowpath_common+0x78/0xd0
> [    0.000000]  [<ffffffff81043fcf>] warn_slowpath_null+0xf/0x20
> [    0.000000]  [<ffffffff8108e806>] __alloc_pages_nodemask+0x1b6/0x730
> [    0.000000]  [<ffffffff810b4e28>] alloc_pages_current+0x78/0xf0
> [    0.000000]  [<ffffffff8108da69>] __get_free_pages+0x9/0x50
> [    0.000000]  [<ffffffff810bb912>] __kmalloc+0x112/0x120
> [    0.000000]  [<ffffffff8146e773>] vgacon_scrollback_startup+0x13/0x70
> [    0.000000]  [<ffffffff811d29b3>] vgacon_startup+0x2a3/0x420
> [    0.000000]  [<ffffffff816fc556>] con_init+0x1b/0x230
> [    0.000000]  [<ffffffff816fba00>] console_init+0x22/0x42
> [    0.000000]  [<ffffffff816d4b8f>] start_kernel+0x240/0x3be
> [    0.000000]  [<ffffffff816d4289>] x86_64_start_reservations+0x99/0xb9
> [    0.000000]  [<ffffffff816d4389>] x86_64_start_kernel+0xe0/0xf2
> [    0.000000] ---[ end trace 4eaa2a86a8e2da22 ]---
> 
> I also got it on 2.6.31.9, but had waited to ask hoping 2.6.32.2 would fix it.
> 
> Attached is my config