Skip to contentSkip to conversation

2.6.18-rc6-mm1

DISCUSSION Magnus Määttä 3 participants 7 messages Sep 9, 2006 – Sep 16, 2006

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

Hi Andrew,

Sorry, forgot to CC lkml when I sent the first one.

Quoted 1 line · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18-rc6/2.6.18-rc6-mm1/

Quoted 2 lines

- autofs4 mounting of NFS is still sick.

I got this oops on my machine when I ran df on another one which have mounted a few NFS shares from my machine. I got 5 more pretty much identical ones within 10 seconds after the first one (haven't seen any more after these though). Also, dmesg is filled with, about a gazillion of these:

[15164.017991] RPC request reserved 9136 but used 9268
[15164.037431] RPC request reserved 9136 but used 9268
[15164.052988] RPC request reserved 9136 but used 9268

Files are also getting corrupted when transfered from my machine, but using my machine as client works fine.

oops here:

Error (regular_file): read_ksyms stat /proc/ksyms failed No modules in ksyms, skipping objects
No ksyms, skipping lsmod

BUG: unable to handle kernel NULL pointer dereference at virtual 
address 00000000
c04ad300
*pde = 00000000
Oops: 0000 [#1]
CPU:    0
EIP:    0060:[<c04ad300>]    Tainted: P      VLI

Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00210212 (2.6.18-rc6-mm1 #1)
eax: 00000000 ebx: e5299000 ecx: 00000000 edx: e8843620
esi: e5299070 edi: ffff84de ebp: e52a0fb0 esp: e52a0f70
ds: 007b es: 007b ss: 0068

Stack: 00200046 eb499aa0 00000001 eb499a84 00000000 e52a0f9c c04eaa1b 
eb499a84
       00000001 e8843620 e529904c e52a0fb0 c012d70b 00000002 ffff84de 
ffff84de
       e52a0fe0 c02784ba e5299000 e52a0fc4 00000000 fffffeff ffffffff 
fffffef8
Call Trace:
 [<c01041bf>] show_trace_log_lvl+0x2f/0x50
 [<c01042a7>] show_stack_log_lvl+0x97/0xc0
 [<c0104532>] show_registers+0x1f2/0x2a0
 [<c01047dd>] die+0x12d/0x240
 [<c011735c>] do_page_fault+0x3ac/0x650
 [<c04eaeef>] error_code+0x3f/0x44
 [<c02784ba>] nfsd+0x18a/0x2b0
 [<c0103fb7>] kernel_thread_helper+0x7/0x10
Code: 89 45 e8 8b 52 28 83 c6 70 89 55 e4 8b 40 04 83 f8 17 0f 86 6d 

04 00 00 8b 5d 08 8b 83 9c 04 00 00 c7 83 a0 04 00 00 01 00 00 00 <8b> 00 89 04 24 e8 06 d4 ca ff c7 46 04 00 00 00 00 89 c1 89 43

Quoted 1 line
Quoted 1 line

EIP; c04ad300 <svc_process+40/6a0> <=====

Trace; c01041bf <show_trace_log_lvl+2f/50>
Trace; c01042a7 <show_stack_log_lvl+97/c0>
Trace; c0104532 <show_registers+1f2/2a0>
Trace; c01047dd <die+12d/240>
Trace; c011735c <do_page_fault+3ac/650>
Trace; c04eaeef <error_code+3f/44>
Trace; c02784ba <nfsd+18a/2b0>
Trace; c0103fb7 <kernel_thread_helper+7/10>

This architecture has variable length instructions, decoding before eip
is unreliable, take these instructions with a pinch of salt.

Code; c04ad2d5 <svc_process+15/6a0>
00000000 <_EIP>:
Code; c04ad2d5 <svc_process+15/6a0>
0: 89 45 e8 mov %eax,0xffffffe8(%ebp)
Code; c04ad2d8 <svc_process+18/6a0>
3: 8b 52 28 mov 0x28(%edx),%edx
Code; c04ad2db <svc_process+1b/6a0>
6: 83 c6 70 add $0x70,%esi
Code; c04ad2de <svc_process+1e/6a0>
9: 89 55 e4 mov %edx,0xffffffe4(%ebp)
Code; c04ad2e1 <svc_process+21/6a0>
c: 8b 40 04 mov 0x4(%eax),%eax
Code; c04ad2e4 <svc_process+24/6a0>
f: 83 f8 17 cmp $0x17,%eax
Code; c04ad2e7 <svc_process+27/6a0>
12: 0f 86 6d 04 00 00 jbe 485 <_EIP+0x485>
Code; c04ad2ed <svc_process+2d/6a0>
18: 8b 5d 08 mov 0x8(%ebp),%ebx
Code; c04ad2f0 <svc_process+30/6a0>
1b: 8b 83 9c 04 00 00 mov 0x49c(%ebx),%eax
Code; c04ad2f6 <svc_process+36/6a0>
21: c7 83 a0 04 00 00 01 movl $0x1,0x4a0(%ebx)
Code; c04ad2fd <svc_process+3d/6a0>
28: 00 00 00

This decode from eip onwards should be reliable

Code; c04ad300 <svc_process+40/6a0>
00000000 <_EIP>:
Code; c04ad300 <svc_process+40/6a0> <=====
0: 8b 00 mov (%eax),%eax <=====
Code; c04ad302 <svc_process+42/6a0>
2: 89 04 24 mov %eax,(%esp)
Code; c04ad305 <svc_process+45/6a0>
5: e8 06 d4 ca ff call ffcad410 <_EIP+0xffcad410>
Code; c04ad30a <svc_process+4a/6a0>
a: c7 46 04 00 00 00 00 movl $0x0,0x4(%esi)
Code; c04ad311 <svc_process+51/6a0>
11: 89 c1 mov %eax,%ecx
Code; c04ad313 <svc_process+53/6a0>
13: 89 .byte 0x89
Code; c04ad314 <svc_process+54/6a0>
14: 43 inc %ebx

EIP: [<c04ad300>] svc_process+0x40/0x6a0 SS:ESP 0068:e52a0f70 Warning (Oops_read): Code line not seen, dumping what data is available

Quoted 1 line
Quoted 1 line

EIP; c04ad300 <svc_process+40/6a0> <=====

2 warnings and 1 error issued. Results may not be reliable.

#1 lore raw
Quoted 15 lines · On Sat, 9 Sep 2006 14:45:32 +0200 Magnus Määttä <[email protected]> wrote:

Hi Andrew,

Sorry, forgot to CC lkml when I sent the first one.

Quoted 1 line · Andrew Morton

ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.18-rc6/2.6.18-rc6-mm1/

Quoted 2 lines

- autofs4 mounting of NFS is still sick.

I got this oops on my machine when I ran df on another one which have mounted a few NFS shares from my machine. I got 5 more pretty much identical ones within 10 seconds after the first one (haven't seen any more after these though). Also, dmesg is filled with, about a gazillion of these:

Quoted 6 lines
[15164.017991] RPC request reserved 9136 but used 9268
[15164.037431] RPC request reserved 9136 but used 9268
[15164.052988] RPC request reserved 9136 but used 9268

Files are also getting corrupted when transfered from my machine, but using my machine as client works fine.

OK, so the NFS server isn't happy.

Quoted 3 lines

oops here:

Error (regular_file): read_ksyms stat /proc/ksyms failed

You don't need to run ksymoops at all in 2.6 - simply enable CONFIG_KALLSYMS and the kernel does the rest.

Quoted 9 lines

No modules in ksyms, skipping objects
No ksyms, skipping lsmod

BUG: unable to handle kernel NULL pointer dereference at virtual 
address 00000000
c04ad300
*pde = 00000000
Oops: 0000 [#1]
CPU:    0
EIP:    0060:[<c04ad300>]    Tainted: P      VLI

What caused the taint?

Quoted 93 lines

Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00210212 (2.6.18-rc6-mm1 #1)
eax: 00000000 ebx: e5299000 ecx: 00000000 edx: e8843620
esi: e5299070 edi: ffff84de ebp: e52a0fb0 esp: e52a0f70
ds: 007b es: 007b ss: 0068

Stack: 00200046 eb499aa0 00000001 eb499a84 00000000 e52a0f9c c04eaa1b 
eb499a84
       00000001 e8843620 e529904c e52a0fb0 c012d70b 00000002 ffff84de 
ffff84de
       e52a0fe0 c02784ba e5299000 e52a0fc4 00000000 fffffeff ffffffff 
fffffef8
Call Trace:
 [<c01041bf>] show_trace_log_lvl+0x2f/0x50
 [<c01042a7>] show_stack_log_lvl+0x97/0xc0
 [<c0104532>] show_registers+0x1f2/0x2a0
 [<c01047dd>] die+0x12d/0x240
 [<c011735c>] do_page_fault+0x3ac/0x650
 [<c04eaeef>] error_code+0x3f/0x44
 [<c02784ba>] nfsd+0x18a/0x2b0
 [<c0103fb7>] kernel_thread_helper+0x7/0x10
Code: 89 45 e8 8b 52 28 83 c6 70 89 55 e4 8b 40 04 83 f8 17 0f 86 6d 

04 00 00 8b 5d 08 8b 83 9c 04 00 00 c7 83 a0 04 00 00 01 00 00 00 <8b> 00 89 04 24 e8 06 d4 ca ff c7 46 04 00 00 00 00 89 c1 89 43

Quoted 1 line
Quoted 1 line

EIP; c04ad300 <svc_process+40/6a0> <=====

Trace; c01041bf <show_trace_log_lvl+2f/50>
Trace; c01042a7 <show_stack_log_lvl+97/c0>
Trace; c0104532 <show_registers+1f2/2a0>
Trace; c01047dd <die+12d/240>
Trace; c011735c <do_page_fault+3ac/650>
Trace; c04eaeef <error_code+3f/44>
Trace; c02784ba <nfsd+18a/2b0>
Trace; c0103fb7 <kernel_thread_helper+7/10>

This architecture has variable length instructions, decoding before eip
is unreliable, take these instructions with a pinch of salt.

Code; c04ad2d5 <svc_process+15/6a0>
00000000 <_EIP>:
Code; c04ad2d5 <svc_process+15/6a0>
0: 89 45 e8 mov %eax,0xffffffe8(%ebp)
Code; c04ad2d8 <svc_process+18/6a0>
3: 8b 52 28 mov 0x28(%edx),%edx
Code; c04ad2db <svc_process+1b/6a0>
6: 83 c6 70 add $0x70,%esi
Code; c04ad2de <svc_process+1e/6a0>
9: 89 55 e4 mov %edx,0xffffffe4(%ebp)
Code; c04ad2e1 <svc_process+21/6a0>
c: 8b 40 04 mov 0x4(%eax),%eax
Code; c04ad2e4 <svc_process+24/6a0>
f: 83 f8 17 cmp $0x17,%eax
Code; c04ad2e7 <svc_process+27/6a0>
12: 0f 86 6d 04 00 00 jbe 485 <_EIP+0x485>
Code; c04ad2ed <svc_process+2d/6a0>
18: 8b 5d 08 mov 0x8(%ebp),%ebx
Code; c04ad2f0 <svc_process+30/6a0>
1b: 8b 83 9c 04 00 00 mov 0x49c(%ebx),%eax
Code; c04ad2f6 <svc_process+36/6a0>
21: c7 83 a0 04 00 00 01 movl $0x1,0x4a0(%ebx)
Code; c04ad2fd <svc_process+3d/6a0>
28: 00 00 00

This decode from eip onwards should be reliable

Code; c04ad300 <svc_process+40/6a0>
00000000 <_EIP>:
Code; c04ad300 <svc_process+40/6a0> <=====
0: 8b 00 mov (%eax),%eax <=====
Code; c04ad302 <svc_process+42/6a0>
2: 89 04 24 mov %eax,(%esp)
Code; c04ad305 <svc_process+45/6a0>
5: e8 06 d4 ca ff call ffcad410 <_EIP+0xffcad410>
Code; c04ad30a <svc_process+4a/6a0>
a: c7 46 04 00 00 00 00 movl $0x0,0x4(%esi)
Code; c04ad311 <svc_process+51/6a0>
11: 89 c1 mov %eax,%ecx
Code; c04ad313 <svc_process+53/6a0>
13: 89 .byte 0x89
Code; c04ad314 <svc_process+54/6a0>
14: 43 inc %ebx

EIP: [<c04ad300>] svc_process+0x40/0x6a0 SS:ESP 0068:e52a0f70 Warning (Oops_read): Code line not seen, dumping what data is available

Quoted 1 line
Quoted 1 line

EIP; c04ad300 <svc_process+40/6a0> <=====

2 warnings and 1 error issued. Results may not be reliable.

#2 lore raw

Hi again!

Quoted 16 lines · Andrew Morton

On Sat, 9 Sep 2006 14:45:32 +0200

Quoted 11 lines · Magnus Määttä

I got this oops on my machine when I ran df on another one which have mounted a few NFS shares from my machine. I got 5 more pretty much identical ones within 10 seconds after the first one (haven't seen any more after these though). Also, dmesg is filled with, about a gazillion of these:

[15164.017991] RPC request reserved 9136 but used 9268
[15164.037431] RPC request reserved 9136 but used 9268
[15164.052988] RPC request reserved 9136 but used 9268

Files are also getting corrupted when transfered from my machine, but using my machine as client works fine.

Quoted 8 lines

OK, so the NFS server isn't happy.

Quoted 3 lines

oops here:

Error (regular_file): read_ksyms stat /proc/ksyms failed

You don't need to run ksymoops at all in 2.6 - simply enable CONFIG_KALLSYMS and the kernel does the rest.

Oh, didn't know that!

Quoted 11 lines
Quoted 9 lines

No modules in ksyms, skipping objects
No ksyms, skipping lsmod

BUG: unable to handle kernel NULL pointer dereference at virtual
address 00000000
c04ad300
*pde = 00000000
Oops: 0000 [#1]
CPU:    0
EIP:    0060:[<c04ad300>]    Tainted: P      VLI

What caused the taint?

I was pretty sure it was listed somewhere, but I guess it wasn't. nvidia graphics module, I can try without it tomorrow if needed.

Ah, it was only listed in the original one:

[  218.012273] BUG: unable to handle kernel NULL pointer dereference at virtual address 00000000
[  218.012289]  printing eip:
[  218.012292] c04ad300
[  218.012294] *pde = 00000000
[  218.012302] Oops: 0000 [#1]
[  218.021009] 4K_STACKS PREEMPT
[  218.030577] last sysfs file: /class/input/input1/name
[  218.046224] Modules linked in: snd_seq_midi snd_seq_oss ipaq usbserial nvidia agpgart eeprom snd_seq_dummy snd_pcm_oss snd_mixer_oss snd_emu10k1_synth snd_emux_synth snd_seq_virmidi snd_seq_midi_event snd_seq_midi_emul snd_seq snd_emu10k1 snd_rawmidi snd_ac97_codec snd_ac97_bus snd_pcm snd_seq_device snd_page_alloc snd_util_mem snd_hwdep psmouse
[  218.142717] CPU:    0
[  218.142718] EIP:    0060:[<c04ad300>]    Tainted: P      VLI
[  218.142720] EFLAGS: 00210212   (2.6.18-rc6-mm1 #1)
[  218.182388] EIP is at svc_process+0x40/0x6a0
[  218.195608] eax: 00000000   ebx: e5299000   ecx: 00000000   edx: e8843620
[  218.216639] esi: e5299070   edi: ffff84de   ebp: e52a0fb0   esp: e52a0f70
[  218.237666] ds: 007b   es: 007b   ss: 0068
[  218.250352] Process nfsd (pid: 5040, ti=e52a0000 task=e52a5570 task.ti=e52a0000)
[  218.272754] Stack: 00200046 eb499aa0 00000001 eb499a84 00000000 e52a0f9c c04eaa1b eb499a84
[  218.299267]        00000001 e8843620 e529904c e52a0fb0 c012d70b 00000002 ffff84de ffff84de
[  218.325781]        e52a0fe0 c02784ba e5299000 e52a0fc4 00000000 fffffeff ffffffff fffffef8
[  218.352294] Call Trace:
[  218.360459]  [<c01041bf>] show_trace_log_lvl+0x2f/0x50
[  218.394276]  [<c01042a7>] show_stack_log_lvl+0x97/0xc0
[  218.428065]  [<c0104532>] show_registers+0x1f2/0x2a0
[  218.461416]  [<c01047dd>] die+0x12d/0x240
[  218.491904]  [<c011735c>] do_page_fault+0x3ac/0x650
[  218.525178]  [<c04eaeef>] error_code+0x3f/0x44
[  218.557331]  [<c02784ba>] nfsd+0x18a/0x2b0
[  218.588522]  [<c0103fb7>] kernel_thread_helper+0x7/0x10
[  218.623405]  =======================
[  218.653244] Code: 89 45 e8 8b 52 28 83 c6 70 89 55 e4 8b 40 04 83 f8 17 0f 86 6d 04 00 00 8b 5d 08 8b 83 9c 04 00 00 c7 83 a0 04 00 00 01 00 00 00 <8b> 00 89 04 24 e8 06 d4 ca ff c7 46 04 00 00 00 00 89 c1 89 43
[  218.755843] EIP: [<c04ad300>] svc_process+0x40/0x6a0 SS:ESP 0068:e52a0f70

Regards,
Magnus Määttä

#3 lore raw
Quoted 7 lines · On Sun, 10 Sep 2006 02:37:51 +0200 Magnus Määttä <[email protected]> wrote:
Quoted 3 lines
Quoted 1 line

EIP: 0060:[<c04ad300>] Tainted: P VLI

What caused the taint?

I was pretty sure it was listed somewhere, but I guess it wasn't. nvidia graphics module, I can try without it tomorrow if needed.

That would be appreciated thanks.

But first you'd need to ensure that it's a repeatable oops. If it's not, the removing the nvidia driver won't tell us if the nvidia driver caused it. (It almost certainly didn't).

#4 lore raw
Quoted 11 lines · Andrew Morton
Quoted 7 lines · On Sun, 10 Sep 2006 02:37:51 +0200 Magnus Määttä <[email protected]> wrote:
Quoted 3 lines
Quoted 1 line

EIP: 0060:[<c04ad300>] Tainted: P VLI

What caused the taint?

I was pretty sure it was listed somewhere, but I guess it wasn't. nvidia graphics module, I can try without it tomorrow if needed.

Quoted 5 lines

That would be appreciated thanks.

But first you'd need to ensure that it's a repeatable oops. If it's not, the removing the nvidia driver won't tell us if the nvidia driver caused it. (It almost certainly didn't).

Forgot to mention that I had this (well, I'm not 100% it was exactly the same) oops with 2.6.18-rc5-mm1.

Here is the oops with an untainted kernel:

[  161.078154] BUG: unable to handle kernel NULL pointer dereference at virtual address 00000000
[  161.104908]  printing eip:
[  161.133157] c04ad300
[  161.158755] *pde = 00000000
[  161.186042] Oops: 0000 [#2]
[  161.212933] 4K_STACKS PREEMPT
[  161.240791] last sysfs file: /class/input/input1/name
[  161.274954] Modules linked in: snd_seq_midi snd_seq_oss ipaq usbserial eeprom snd_seq_dummy snd_pcm_oss snd_mixer_oss snd_emu10k1_synth snd_emux_synth snd_seq_virmidi snd_seq_midi_event snd_seq_midi_emul snd_seq snd_emu10k1 snd_rawmidi snd_ac97_codec snd_ac97_bus snd_pcm snd_seq_device snd_page_alloc snd_util_mem snd_hwdep psmouse
[  161.428307] CPU:    0
[  161.428308] EIP:    0060:[<c04ad300>]    Not tainted VLI
[  161.428310] EFLAGS: 00010216   (2.6.18-rc6-mm1 #1)
[  161.531660] EIP is at svc_process+0x40/0x6a0
[  161.566957] eax: 00000000   ebx: eaf8c000   ecx: 00000000   edx: ead74d40
[  161.610544] esi: eaf8c070   edi: ffff5225   ebp: eaf93fb0   esp: eaf93f70
[  161.654260] ds: 007b   es: 007b   ss: 0068
[  161.689932] Process nfsd (pid: 4655, ti=eaf93000 task=eaf9caa0 task.ti=eaf93000)
[  161.712308] Stack: 00000046 eeedefe0 00000001 eeedefc4 00000000 eaf93f9c c04eaa1b eeedefc4
[  161.762411]        00000001 ead74d40 eaf8c04c eaf93fb0 c012d70b 00000000 ffff5226 ffff5225
[  161.812643]        eaf93fe0 c02784ba eaf8c000 eaf93fc4 00000000 fffffeff ffffffff fffffef8
[  161.862772] Call Trace:
[  161.917732]  [<c01041bf>] show_trace_log_lvl+0x2f/0x50
[  161.956879]  [<c01042a7>] show_stack_log_lvl+0x97/0xc0
[  161.995482]  [<c0104532>] show_registers+0x1f2/0x2a0
[  162.033228]  [<c01047dd>] die+0x12d/0x240
[  162.067860]  [<c011735c>] do_page_fault+0x3ac/0x650
[  162.104983]  [<c04eaeef>] error_code+0x3f/0x44
[  162.140705]  [<c02784ba>] nfsd+0x18a/0x2b0
[  162.175232]  [<c0103fb7>] kernel_thread_helper+0x7/0x10
[  162.213210]  =======================
[  162.246050] Code: 89 45 e8 8b 52 28 83 c6 70 89 55 e4 8b 40 04 83 f8 17 0f 86 6d 04 00 00 8b 5d 08 8b 83 9c 04 00 00 c7 83 a0 04 00 00 01 00 00 00 <8b> 00 89 04 24 e8 06 d4 ca ff c7 46 04 00 00 00 00 89 c1 89 43
[  162.354045] EIP: [<c04ad300>] svc_process+0x40/0x6a0 SS:ESP 0068:eaf93f70

And oops #1 just in case it's not unrelated (and the BUG just after it):

[  105.928488] BUG: unable to handle kernel paging request at virtual address b79505a0
[  105.928503]  printing eip:
[  105.928506] c013d14c
[  105.928509] *pde = 2d268067
[  105.928511] *pte = 00000000
[  105.928518] Oops: 0000 [#1]
[  105.952628] 4K_STACKS PREEMPT
[  105.977498] last sysfs file: /class/input/input1/name
[  106.008548] Modules linked in: snd_seq_midi snd_seq_oss ipaq usbserial eeprom snd_seq_dummy snd_pcm_oss snd_mixer_oss snd_emu10k1_synth snd_emux_synth snd_seq_virmidi snd_seq_midi_event snd_seq_midi_emul snd_seq snd_emu10k1 snd_rawmidi snd_ac97_codec snd_ac97_bus snd_pcm snd_seq_device snd_page_alloc snd_util_mem snd_hwdep psmouse
[  106.150657] CPU:    0
[  106.150658] EIP:    0060:[<c013d14c>]    Not tainted VLI
[  106.150660] EFLAGS: 00010002   (2.6.18-rc6-mm1 #1)
[  106.241322] EIP is at trace_hardirqs_on+0x1c/0x150
[  106.273081] eax: 00000001   ebx: b794fba0   ecx: b794f2e0   edx: b7f61410
[  106.310907] esi: 00000008   edi: b7c08ff4   ebp: eee0dfa4   esp: eee0df8c
[  106.348781] ds: 007b   es: 007b   ss: 0068
[  106.378378] Process icecast (pid: 4446, ti=eee0d000 task=ee9d0050 task.ti=eee0d000)
[  106.401560] Stack: eee0df9c 00000046 00000000 00000000 00000000 00000008 00000000 c01033ba
[  106.445119]        b7f61410 b794f2e0 00000000 00000000 b794f3b0 00000000 00000008 b7c08ff4
[  106.489043]        b794f388 00000000 0000007b 0000007b 00000033 000000af b7f61410 00000073
[  106.533124] Call Trace:
[  106.576870]  [<c01041bf>] show_trace_log_lvl+0x2f/0x50
[  106.610486]  [<c01042a7>] show_stack_log_lvl+0x97/0xc0
[  106.643976]  [<c0104532>] show_registers+0x1f2/0x2a0
[  106.676895]  [<c01047dd>] die+0x12d/0x240
[  106.707110]  [<c011735c>] do_page_fault+0x3ac/0x650
[  106.740133]  [<c04eaeef>] error_code+0x3f/0x44
[  106.771908]  [<c01033ba>] sysenter_past_esp+0x93/0x99
[  106.805629]  =======================
[  106.834887] Code: c7 05 94 9a 8b c0 01 00 00 00 89 e5 c9 c3 90 55 89 e5 56 53 83 ec 10 a1 a0 29 63 c0 65 8b 1d 00 00 00 00 85 c0 0f 84 89 00 00 00 <8b> b3 00 0a 00 00 85 f6 75 7f 8b 0d 94 9a 8b c0 85 c9 0f 84 96
[  106.978502]  <3>BUG: sleeping function called from invalid context at kernel/rwsem.c:20
[  107.023854] in_atomic():0, irqs_disabled():1
[  107.058071]  [<c01041bf>] show_trace_log_lvl+0x2f/0x50
[  107.095140]  [<c0104207>] show_trace+0x27/0x30
[  107.129851]  [<c0104334>] dump_stack+0x24/0x30
[  107.164377]  [<c011d442>] __might_sleep+0xa2/0xc0
[  107.199995]  [<c013949e>] down_read+0x1e/0x60
[  107.234807]  [<c012e9b7>] blocking_notifier_call_chain+0x17/0x50
[  107.275150]  [<c0121881>] profile_task_exit+0x21/0x30
[  107.312713]  [<c012327b>] do_exit+0x1b/0x520
[  107.348071]  [<c01048e5>] die+0x235/0x240
[  107.382754]  [<c011735c>] do_page_fault+0x3ac/0x650
[  107.419931]  [<c04eaeef>] error_code+0x3f/0x44
[  107.455575]  [<c01033ba>] sysenter_past_esp+0x93/0x99
[  107.492825]  =======================
#5 lore raw
In reply to Andrew Morton
Quoted 6 lines · On Saturday September 9, [email protected] wrote:
Quoted 4 lines · On Sat, 9 Sep 2006 14:45:32 +0200 Magnus Määttä <[email protected]> wrote:
[15164.017991] RPC request reserved 9136 but used 9268
[15164.037431] RPC request reserved 9136 but used 9268
[15164.052988] RPC request reserved 9136 but used 9268

Don't know what is causing this yet....

Quoted 3 lines
Quoted 3 lines

Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00210212 (2.6.18-rc6-mm1 #1)
eax: 00000000 ebx: e5299000 ecx: 00000000 edx: e8843620

..

Quoted 8 lines
Quoted 8 lines
 [<c02784ba>] nfsd+0x18a/0x2b0
 [<c0103fb7>] kernel_thread_helper+0x7/0x10
Code: 89 45 e8 8b 52 28 83 c6 70 89 55 e4 8b 40 04 83 f8 17 0f 86 6d 

04 00 00 8b 5d 08 8b 83 9c 04 00 00 c7 83 a0 04 00 00 01 00 00 00 <8b> 00 89 04 24 e8 06 d4 ca ff c7 46 04 00 00 00 00 89 c1 89 43

Quoted 1 line
Quoted 1 line

EIP; c04ad300 <svc_process+40/6a0> <=====

But this is probably fixed by the following patch. Can you confirm?

thanks,
NeilBrown

### Diffstat output

 ./net/sunrpc/svcsock.c |    1 +
 1 file changed, 1 insertion(+)
net/sunrpc/svcsock.c +1 -0
diff .prev/net/sunrpc/svcsock.c ./net/sunrpc/svcsock.c
@@ -1709,6 +1709,7 @@ static int svc_deferred_recv(struct svc_
 	rqstp->rq_prot        = dr->prot;
 	rqstp->rq_addr        = dr->addr;
 	rqstp->rq_daddr       = dr->daddr;
+	rqstp->rq_respages    = rqstp->rq_pages;
 	return dr->argslen<<2;
 }
 
#6 lore raw

Hi,

Sorry for the late reply, been away a few days.

Quoted 21 lines · Neil Brown
Quoted 6 lines · On Saturday September 9, [email protected] wrote:
Quoted 4 lines · On Sat, 9 Sep 2006 14:45:32 +0200 Magnus Määttä <[email protected]> wrote:
[15164.017991] RPC request reserved 9136 but used 9268
[15164.037431] RPC request reserved 9136 but used 9268
[15164.052988] RPC request reserved 9136 but used 9268

Don't know what is causing this yet....

Quoted 3 lines
Quoted 3 lines

Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00210212 (2.6.18-rc6-mm1 #1)
eax: 00000000 ebx: e5299000 ecx: 00000000 edx: e8843620

..

Quoted 7 lines
Quoted 7 lines
 [<c02784ba>] nfsd+0x18a/0x2b0
 [<c0103fb7>] kernel_thread_helper+0x7/0x10
Code: 89 45 e8 8b 52 28 83 c6 70 89 55 e4 8b 40 04 83 f8 17 0f 86 6d 

04 00 00 8b 5d 08 8b 83 9c 04 00 00 c7 83 a0 04 00 00 01 00 00 00 <8b> 00 89 04 24 e8 06 d4 ca ff c7 46 04 00 00 00 00 89 c1 89 43

Quoted 4 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

EIP; c04ad300 <svc_process+40/6a0> <=====

But this is probably fixed by the following patch. Can you confirm?

Yes, the patch fixed it, thanks!

Regards,
Magnus