Skip to contentSkip to conversation

Ooops in 2.4.18 through 2.4.20, now kswapd is defunct

DISCUSSION Zhenghui Zhou 1 participant 1 message Mar 27, 2003

lore
Share & tools
lore
#0 lore raw
Quoted 8 lines

I have a variety of systems running kernels ranging from 2.4.18 through 2.4.20 and am seeing fairly frequent kernel oopsen. After the oops, kswapd is defunct, however the systems are still running (they are dual CPU systems). Has anyone seen this before, and is there a patch to fix it yet ? Thanks.

Quoted 7 lines

Kelvin Edwards
System Admin
Jefferson Lab

Here's some Ooops run through ksymoops:

Mar 21 15:20:00 MachX kernel: Unable to handle kernel NULL pointer

dereference

Quoted 6 lines

at virtual address 00000004

Mar 21 15:20:00 MachX kernel: c0152b51
Mar 21 15:20:00 MachX kernel: *pde = 00000000
Mar 21 15:20:00 MachX kernel: Oops: 0000
Mar 21 15:20:00 MachX kernel: CPU:    4
Mar 21 15:20:00 MachX kernel: EIP:    0010:[destroy_inode+33/80]

Not

Quoted 5 lines

tainted

Mar 21 15:20:00 MachX kernel: EIP:    0010:[<c0152b51>]    Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
Mar 21 15:20:00 MachX kernel: EFLAGS: 00010246
Mar 21 15:20:00 MachX kernel: eax: 00000000   ebx: e3c999c0   ecx:

00000000

Quoted 2 lines

edx: e3c999c0

Mar 21 15:20:00 MachX kernel: esi: e3c999c0   edi: 00000001   ebp:

00000a55

Quoted 3 lines

esp: f7a8fefc

Mar 21 15:20:00 MachX kernel: ds: 0018   es: 0018   ss: 0018
Mar 21 15:20:00 MachX kernel: Process kswapd (pid: 11,

stackpage=f7a8f000)

Quoted 1 line
Mar 21 15:20:00 MachX kernel: Stack: e3c999c0 c0154335 e3c999c0

d940e7a0

Quoted 2 lines

c034c700 00150c00 f8a09e9a ca60c738

Mar 21 15:20:00 MachX kernel:        ca60c720 e3c999c0 c0151a31

e3c999c0

Quoted 2 lines

e3c999c0 c0135e63 d7f4e400 f7a8e000

Mar 21 15:20:00 MachX kernel:        ffffffff 000001d0 c02e6308

f7a8e000

Quoted 4 lines

00000003 0000001f 000001d0 00000006

Mar 21 15:20:00 MachX kernel: Call Trace:    [iput+629/640]

[appletalk:__insmod_appletalk_S.bss_L268+425722/117747542] [prune_dcache+225/368] [shrink_cache+819/976]

[shrink_dcache_memory+32/48]

Quoted 6 lines
Mar 21 15:20:00 MachX kernel: Call Trace:    [<c0154335>] [<f8a09e9a>]
[<c0151a31>] [<c0135e63>] [<c0151de0>]
Mar 21 15:20:00 MachX kernel:   [<c0136087>] [<c01360ec>] [<c01361ff>]
[<c0136276>] [<c01363b1>] [<c0136310>]
Mar 21 15:20:00 MachX kernel:   [<c0105000>] [<c0107296>] [<c0136310>]
Mar 21 15:20:00 MachX kernel: Code: 8b 40 04 85 c0 74 08 53 ff d0 59

eb 11 89

Quoted 33 lines

f6 53 8b 15 b4 ae

Quoted 1 line
Quoted 1 line

EIP; c0152b51 <destroy_inode+21/50> <=====

Trace; c0154335 <iput+275/280>
Trace; f8a09e9a <[nfs]nfs_dentry_iput+5a/80> Trace; c0151a31 <prune_dcache+e1/170>
Trace; c0135e63 <shrink_cache+333/3d0>
Trace; c0151de0 <shrink_dcache_memory+20/30> Trace; c0136087 <shrink_caches+67/90>
Trace; c01360ec <try_to_free_pages_zone+3c/60> Trace; c01361ff <kswapd_balance_pgdat+4f/a0> Trace; c0136276 <kswapd_balance+26/40>
Trace; c01363b1 <kswapd+a1/ba>
Trace; c0136310 <kswapd+0/ba>
Trace; c0105000 <_stext+0/0>
Trace; c0107296 <kernel_thread+26/30>
Trace; c0136310 <kswapd+0/ba>
Code; c0152b51 <destroy_inode+21/50>
00000000 <_EIP>:
Code; c0152b51 <destroy_inode+21/50> <=====
0: 8b 40 04 mov 0x4(%eax),%eax <=====
Code; c0152b54 <destroy_inode+24/50>
3: 85 c0 test %eax,%eax
Code; c0152b56 <destroy_inode+26/50>
5: 74 08 je f <_EIP+0xf> c0152b60
<destroy_inode+30/50>
Code; c0152b58 <destroy_inode+28/50>
7: 53 push %ebx
Code; c0152b59 <destroy_inode+29/50>
8: ff d0 call *%eax
Code; c0152b5b <destroy_inode+2b/50>
a: 59 pop %ecx
Code; c0152b5c <destroy_inode+2c/50>

Quoted 9 lines

b: eb 11 jmp 1e <_EIP+0x1e> c0152b6f
<destroy_inode+3f/50>
Code; c0152b5e <destroy_inode+2e/50>
d: 89 f6 mov %esi,%esi
Code; c0152b60 <destroy_inode+30/50>
f: 53 push %ebx
Code; c0152b61 <destroy_inode+31/50>
10: 8b 15 b4 ae 00 00 mov 0xaeb4,%edx

I meet the similar situation, I run the server on internet with heavy stress and cannot trace it clearly, I also tested from 2.4.18 to 2.4.20, and got the same wrong thing, I have to limit the number of processes running on the server to cut down the errors.

The error fired while load and run a program from disk, if it is do by hand, it shows as "Segmentation Fault". The dmesg shows:

<1>Unable to handle kernel NULL pointer dereference at virtual address 00000004

	 printing eip:
	dfd91718
	*pde = 00000000
	Oops: 0000
	CPU:    0
	EIP:    0010:[<dfd91718>]    Not tainted
	EFLAGS: 00010286
	eax: bffffae4   ebx: cff46000   ecx: 00000000   edx: cff46000
	esi: c0106c33   edi: 0000000b   ebp: cff47fb8   esp: cff47f84
	ds: 0018   es: 0018   ss: 0018
	Process more (pid: 20846, stackpage=cff47000)
	Stack: cff46000 c0106c33 0000000b 00000000 d3ef2000 0000000b
cff47fbc c0105987
     	  bffffae4 c01059a7 00000000 00000a3a 00000020 bffffa5c dfd918e8
00000000
	       00000000 00000000 00000000 00000000 00000000 00000000
00000000 0000002b
	Call Trace:    [<c0106c33>] [<c0105987>] [<c01059a7>]
	
	Code: 8b 51 04 83 fa ff 0f 84 56 01 00 00 83 fa fc 77 07 c7 41

04

I run ksymoops with correct specified vmlinux and System.map, the result shows:

Warning (compare_maps): ksyms_base symbol
default_idle_R__ver_default_idle not f
ound in vmlinux. Ignoring ksyms_base entry Warning (compare_maps): ksyms_base symbol
machine_real_restart_R__ver_machine_re
al_restart not found in vmlinux. Ignoring ksyms_base entry Reading Oops report from the terminal
Using defaults from ksymoops -t elf32-i386 -a i386

<1>Unable to handle kernel NULL pointer dereference at virtual
address 00000004

	dfd91718
	*pde = 00000000
	Oops: 0000
	CPU:    0
	EIP:    0010:[<dfd91718>]    Not tainted
	EFLAGS: 00010286
	eax: bffffae4   ebx: cff46000   ecx: 00000000   edx: cff46000
	esi: c0106c33   edi: 0000000b   ebp: cff47fb8   esp: cff47f84
	ds: 0018   es: 0018   ss: 0018
	Process more (pid: 20846, stackpage=cff47000)
	Stack: cff46000 c0106c33 0000000b 00000000 d3ef2000 0000000b
cff47fbc c0105987
	       bffffae4 c01059a7 00000000 00000a3a 00000020 bffffa5c
dfd918e8 00000000
	       00000000 00000000 00000000 00000000 00000000 00000000
00000000 0000002b
	Call Trace:    [<c0106c33>] [<c0105987>] [<c01059a7>]
	Code: 8b 51 04 83 fa ff 0f 84 56 01 00 00 83 fa fc 77 07 c7 41
04
	>>EIP; dfd91718 <_end+1fb00020/205fa908>   <=====
	
	>>eax; bffffae4 Before first symbol
	>>ebx; cff46000 <_end+fcb4908/205fa908>
	>>edx; cff46000 <_end+fcb4908/205fa908>
	>>esi; c0106c33 <system_call+2f/34>
	>>ebp; cff47fb8 <_end+fcb68c0/205fa908>
	>>esp; cff47f84 <_end+fcb688c/205fa908>
	
	Trace; c0106c33 <system_call+2f/34>
	Trace; c0105987 <sys_execve+2f/60>
	Trace; c01059a7 <sys_execve+4f/60>
	
	Code;  dfd91718 <_end+1fb00020/205fa908>
	00000000 <_EIP>:
	Code;  dfd91718 <_end+1fb00020/205fa908>   <=====
	   0:   8b 51 04                  mov    0x4(%ecx),%edx   <=====
	Code;  dfd9171b <_end+1fb00023/205fa908>
	   3:   83 fa ff                  cmp    $0xffffffff,%edx
	Code;  dfd9171e <_end+1fb00026/205fa908>
	   6:   0f 84 56 01 00 00         je     162 <_EIP+0x162>
dfd9187a <_end+1fb0018
	2/205fa908>
	Code;  dfd91724 <_end+1fb0002c/205fa908>
	   c:   83 fa fc                  cmp    $0xfffffffc,%edx
	Code;  dfd91727 <_end+1fb0002f/205fa908>
	   f:   77 07                     ja     18 <_EIP+0x18> dfd91730
<_end+1fb00038/
	205fa908>
	Code;  dfd91729 <_end+1fb00031/205fa908>
	  11:   c7 41 04 00 00 00 00      movl   $0x0,0x4(%ecx)

Linux jsgx 2.4.20 #2 Tue Mar 25 22:47:46 CST 2003 i686 unknown

Gnu C                  2.96
Gnu make               3.78.1
binutils               2.11.90.0.8
util-linux             2.10f
mount                  2.10r
modutils               2.4.13
e2fsprogs              1.32
pcmcia-cs              3.1.20
PPP                    2.4.1
isdn4k-utils           3.1beta7
Linux C Library        so.8.0
Linux C Library        2.1.3
Dynamic linker (ldd)   2.1.3
Procps                 2.0.6
Net-tools              1.60
Console-tools          0.3.3
Sh-utils               2.0

Modules Loaded

I tested several situations, cann't get rid of it.