[OOPS] kswapd 2.4.20

[OOPS] kswapd 2.4.20

[David Lewis]

Fourth one sent to list.. Please advise if any other information is needed.
A reply would be appreciated even if it is to say that this is the wrong
place.

May 21 03:23:50 nicebox kernel: Unable to handle kernel paging request at
virtual address 938a909e
May 21 03:23:50 nicebox kernel: c013670c
May 21 03:23:50 nicebox kernel: *pde = 00000000
May 21 03:23:50 nicebox kernel: Oops: 0002
May 21 03:23:50 nicebox kernel: CPU:    1
May 21 03:23:50 nicebox kernel: EIP:    0010:[<c013670c>]    Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
May 21 03:23:50 nicebox kernel: EFLAGS: 00210046
May 21 03:23:50 nicebox kernel: eax: 938a909a   ebx: d7d0b000   ecx:
0000000a   edx: da851000
May 21 03:23:50 nicebox kernel: esi: c1577268   edi: c1577268   ebp:
00000954   esp: df6d5f08
May 21 03:23:50 nicebox kernel: ds: 0018   es: 0018   ss: 0018
May 21 03:23:50 nicebox kernel: Process kswapd (pid: 5, stackpage=df6d5000)
May 21 03:23:50 nicebox kernel: Stack: 0000003d df6df304 c01359b2 c1577268
d7d0b600 df6df000 00200292 da687900
May 21 03:23:50 nicebox kernel:        c0135a94 c1577268 df6df200 0000007e
c602ab80 c67f8100 c602ab80 c015507a
May 21 03:23:50 nicebox kernel:        c1577268 c602ab80 0000000c 000001d0
00000020 00000006 c0155444 0000169c
May 21 03:23:50 nicebox kernel: Call Trace:    [<c01359b2>] [<c0135a94>]
[<c015507a>] [<c0155444>] [<c01370a7>]
May 21 03:23:50 nicebox kernel:   [<c0137106>] [<c013721c>] [<c0137298>]
[<c01373cd>] [<c0105000>] [<c01058ce>]
May 21 03:23:50 nicebox kernel:   [<c0137330>]
May 21 03:23:50 nicebox kernel: Code: 89 50 04 8b 46 08 8d 56 08 89 58 04 89
03 89 53 04 89 5e 08


>>EIP; c013670c <kmem_cache_free_one+6c/a5>   <=====

>>eax; 938a909a Before first symbol
>>ebx; d7d0b000 <_end+179a4a28/1fea4a88>
>>edx; da851000 <_end+1a4eaa28/1fea4a88>
>>esi; c1577268 <_end+1210c90/1fea4a88>
>>edi; c1577268 <_end+1210c90/1fea4a88>
>>ebp; 00000954 Before first symbol
>>esp; df6d5f08 <_end+1f36f930/1fea4a88>

Trace; c01359b2 <free_block+32/50>
Trace; c0135a94 <kmem_cache_free+64/90>
Trace; c015507a <prune_dcache+10a/1a0>
Trace; c0155444 <shrink_dcache_memory+24/40>
Trace; c01370a7 <shrink_caches+77/a0>
Trace; c0137106 <try_to_free_pages_zone+36/50>
Trace; c013721c <kswapd_balance_pgdat+5c/b0>
Trace; c0137298 <kswapd_balance+28/40>
Trace; c01373cd <kswapd+9d/c0>
Trace; c0105000 <_stext+0/0>
Trace; c01058ce <kernel_thread+2e/40>
Trace; c0137330 <kswapd+0/c0>

Code;  c013670c <kmem_cache_free_one+6c/a5>
00000000 <_EIP>:
Code;  c013670c <kmem_cache_free_one+6c/a5>   <=====
   0:   89 50 04                  mov    %edx,0x4(%eax)   <=====
Code;  c013670f <kmem_cache_free_one+6f/a5>
   3:   8b 46 08                  mov    0x8(%esi),%eax
Code;  c0136712 <kmem_cache_free_one+72/a5>
   6:   8d 56 08                  lea    0x8(%esi),%edx
Code;  c0136715 <kmem_cache_free_one+75/a5>
   9:   89 58 04                  mov    %ebx,0x4(%eax)
Code;  c0136718 <kmem_cache_free_one+78/a5>
   c:   89 03                     mov    %eax,(%ebx)
Code;  c013671a <kmem_cache_free_one+7a/a5>
   e:   89 53 04                  mov    %edx,0x4(%ebx)
Code;  c013671d <kmem_cache_free_one+7d/a5>
  11:   89 5e 08                  mov    %ebx,0x8(%esi)

David Lewis
Senior Security Engineer
VNX Solutions, Inc
dlewis@vnxsolutions.com
410-459-7428 Cell

Re: [OOPS] kswapd 2.4.20

[Marc-Christian Petersen]

On Wednesday 21 May 2003 16:45, David Lewis wrote:

Hi David,

> Fourth one sent to list.. Please advise if any other information is needed.
> A reply would be appreciated even if it is to say that this is the wrong
> place.
> May 21 03:23:50 nicebox kernel: Unable to handle kernel paging request at
> virtual address 938a909e
> May 21 03:23:50 nicebox kernel: Process kswapd (pid: 5, stackpage=df6d5000)
I really don't know how often I've seen this in recent 2.4 kernels 
(.17,.18,.19,.20 (mainstream) and so on).

For me, I use the rmap VM. The problem is eliminated in that tree. Also the 
-aa VM does not have that problem.

I am pretty sure Andrea sent a fix for this issue more than once a while ago 
to the list cc'ed Marcelo (correct me if I am wrong), but to quote Andrea:
"We are here to fix bugs. If bug fixes are not accepted we can stop right here 
to fix bugs".

ciao, Marc

RE: [OOPS] kswapd 2.4.20

[David Lewis]

Hiya Marc,

Thanks for the pointers to the new VM systems. I changed the kernel over to
2.4.21-rc2-ac2 (has rmap in it, right?) and I am rerunning my tests to see
if that solves the problem.

I agree with Andrea about bugs.. I certainly accept them as a necessary
evil. I tend to get frustrated when faced with something that I dont have
the skill to solve myself however. Especially so when I cant seem to get
pointed in the right direction by anyone. :)

Thanks again for your suggestion, and I will let you know if it works out!

David Lewis
Senior Security Engineer
VNX Solutions, Inc <http://www.vnxsolutions.com>
dlewis@vnxsolutions.com <mailto:dlewis@vnxsolutions.com>
410-459-7428 Cell



-----Original Message-----
From: Marc-Christian Petersen [mailto:m.c.p@wolk-project.de]
Sent: Thursday, May 22, 2003 4:41 AM
To: David Lewis; linux-kernel@vger.kernel.org
Cc: Erin Britz
Subject: Re: [OOPS] kswapd 2.4.20


On Wednesday 21 May 2003 16:45, David Lewis wrote:

Hi David,

> Fourth one sent to list.. Please advise if any other information is
needed.
> A reply would be appreciated even if it is to say that this is the wrong
> place.
> May 21 03:23:50 nicebox kernel: Unable to handle kernel paging request at
> virtual address 938a909e
> May 21 03:23:50 nicebox kernel: Process kswapd (pid: 5,
stackpage=df6d5000)
I really don't know how often I've seen this in recent 2.4 kernels
(.17,.18,.19,.20 (mainstream) and so on).

For me, I use the rmap VM. The problem is eliminated in that tree. Also the
-aa VM does not have that problem.

I am pretty sure Andrea sent a fix for this issue more than once a while ago
to the list cc'ed Marcelo (correct me if I am wrong), but to quote Andrea:
"We are here to fix bugs. If bug fixes are not accepted we can stop right
here
to fix bugs".

ciao, Marc

Re: [OOPS] kswapd 2.4.20

[Carl-Daniel Hailfinger]

David Lewis wrote:
> Hiya Marc,
> 
> Thanks for the pointers to the new VM systems. I changed the kernel over to
> 2.4.21-rc2-ac2 (has rmap in it, right?)

No. IIRC, rmap was dropped from -ac

Carl-Daniel
-- 
http://www.hailfinger.org/

Re: [OOPS] kswapd 2.4.20

[Marc-Christian Petersen]

On Thursday 22 May 2003 18:01, David Lewis wrote:

Hi David,

> Thanks for the pointers to the new VM systems. I changed the kernel over to
np.

> 2.4.21-rc2-ac2 (has rmap in it, right?) and I am rerunning my tests to see
> if that solves the problem.
Nope, -ac dropped rmap some weeks ago (or even months). Carldani is right :)

> I agree with Andrea about bugs.. I certainly accept them as a necessary
> evil. I tend to get frustrated when faced with something that I dont have
> the skill to solve myself however. Especially so when I cant seem to get
> pointed in the right direction by anyone. :)
hehe :) ... Well, some people including me are frustrated about sending 
patches (bugfixes) again and again to only see silent ignores, thus also 
frustrated about telling the issues, pointing to fixes etc.

> Thanks again for your suggestion, and I will let you know if it works out!
thanks :) Would be great to hear from you if it'll work (I'll bet it will ;)

ciao, Marc