Reiser4 oops with 2.6.13-rc6-mm2

Reiser4 oops with 2.6.13-rc6-mm2

[Ronald Moesbergen]

Hi,

While running a vmware virtual machine from a reiser4 filesystem, my
machine consistently locks up and produces the following OOPS:

kernel: Oops: 0002 [#1]
kernel: PREEMPT SMP
kernel: last sysfs file: /class/vc/vcsa8/dev
kernel: Modules linked in: vmnet vmmon nvidia
kernel: CPU:    0
kernel: EIP:    0060:[lock_object+84/127]    Tainted: P      VLI
kernel: EIP:    0060:[<b01a1317>]    Tainted: P      VLI
kernel: EFLAGS: 00013202   (2.6.13-rc6-mm2)
kernel: EIP is at lock_object+0x54/0x7f
kernel: eax: cf569e94   ebx: 00000000   ecx: e256fe80   edx: e256fe94
kernel: esi: dba35280   edi: e256fec8   ebp: e256fec8   esp: e256fb48
kernel: ds: 007b   es: 007b   ss: 0068
kernel: Process vmware-vmx (pid: 11869, threadinfo=e256e000 task=eec0c030)
kernel: Stack: dba35280 00000000 e256fec8 b01a16e7 e256fec8 dba352b4
00000000 dba352f0
kernel:        b01a17ba e256fec8 00000000 00000000 00000001 00000000
00000000 e256fec8
kernel:        dba35280 dba352f0 b01a18ad e256fec8 b02a7417 efe1d76c
d9fa46c4 00000016
kernel: Call Trace:
kernel:  [lock_tail+80/113] lock_tail+0x50/0x71
kernel:  [<b01a16e7>] lock_tail+0x50/0x71
kernel:  [longterm_lock_tryfast+178/229] longterm_lock_tryfast+0xb2/0xe5
kernel:  [<b01a17ba>] longterm_lock_tryfast+0xb2/0xe5
kernel:  [longterm_lock_znode+192/695] longterm_lock_znode+0xc0/0x2b7
kernel:  [<b01a18ad>] longterm_lock_znode+0xc0/0x2b7
kernel:  [ata_scsi_translate+265/320] ata_scsi_translate+0x109/0x140
kernel:  [<b02a7417>] ata_scsi_translate+0x109/0x140
kernel:  [ata_scsi_rw_xlat+0/974] ata_scsi_rw_xlat+0x0/0x3ce
kernel:  [<b02a6e95>] ata_scsi_rw_xlat+0x0/0x3ce
kernel:  [cbk_cache_scan_slots+320/754] cbk_cache_scan_slots+0x140/0x2f2
kernel:  [schedule+1320/3501] schedule+0x528/0xdad
kernel:  [<b039752c>] schedule+0x528/0xdad
kernel:  [cbk_cache_search+39/92] cbk_cache_search+0x27/0x5c
kernel:  [<b01b2466>] cbk_cache_search+0x27/0x5c
kernel:  [coord_by_handle+16/41] coord_by_handle+0x10/0x29
kernel:  [<b01b110f>] coord_by_handle+0x10/0x29
kernel:  [object_lookup+188/230] object_lookup+0xbc/0xe6
kernel:  [<b01b10d5>] object_lookup+0xbc/0xe6
kernel:  [find_file_item+322/482] find_file_item+0x142/0x1e2
kernel:  [<b01e4d71>] find_file_item+0x142/0x1e2
kernel:  [read_file+185/530] read_file+0xb9/0x212
kernel:  [<b01e6d83>] read_file+0xb9/0x212
kernel:  [txn_begin+29/33] txn_begin+0x1d/0x21
kernel:  [<b01a74b1>] txn_begin+0x1d/0x21
kernel:  [read_unix_file+532/831] read_unix_file+0x214/0x33f
kernel:  [<b01e70fe>] read_unix_file+0x214/0x33f
kernel:  [init_context+117/189] init_context+0x75/0xbd
kernel:  [<b01a47dd>] init_context+0x75/0xbd
kernel:  [reiser4_read+143/255] reiser4_read+0x8f/0xff
kernel:  [<b01baade>] reiser4_read+0x8f/0xff
kernel:  [vfs_read+328/333] vfs_read+0x148/0x14d
kernel:  [<b015e7e6>] vfs_read+0x148/0x14d
kernel:  [sys_pread64+112/125] sys_pread64+0x70/0x7d
kernel:  [<b015ebd4>] sys_pread64+0x70/0x7d
kernel:  [sysenter_past_esp+84/117] sysenter_past_esp+0x54/0x75
kernel:  [<b0102d9b>] sysenter_past_esp+0x54/0x75
kernel: Code: 04 89 71 08 8d 51 0c 8b 58 04 89 41 0c 89 50 04 89 13 89
5a 04 8d 51 14 83 47 18 01 8b 86 84 00 00 00$
kernel:  <3>BUG: vmware-vmx[11869] exited with nonzero preempt_count 1!

If I move the virtual harddisk images to another partition that
doesn't use reiser4, it runs fine and there are no problems. I suspect
this is a bug in reiser4. Could you assist me in solving this problem?
I'm running kernel 2.6.13-rc6-mm2, but I've also tried 2.6.12 with the
reiser4 patches from your ftp site, with the same effect. Vmware
any-any-93 patches are applied. If you need any more info, please ask.

Thanks,
Ronald.

Re: Reiser4 oops with 2.6.13-rc6-mm2

[evilninja]

Ronald Moesbergen schrieb:
> Hi,
> 
> While running a vmware virtual machine from a reiser4 filesystem, my
> machine consistently locks up and produces the following OOPS:
> 
> kernel: Oops: 0002 [#1]
> kernel: PREEMPT SMP
> kernel: last sysfs file: /class/vc/vcsa8/dev
> kernel: Modules linked in: vmnet vmmon nvidia

is this reproducible and is it reproducible when you don't run this 
inside vmware? the tainted kernel could make it hard to find a solution 
here...

Christian.

Re: Reiser4 oops with 2.6.13-rc6-mm2

[Ronald Moesbergen]

Hi,

Thanks for responding. To clarify: I'm not running linux inside
vmware, I'm running vmware on linux and the vmware images are on a
reiser4 partition. I'm running windows inside vmware. I can certainly
reproduce this, but haven't tried without the nvidia module. I will do
that right now and report back.

Thanks,
Ronald.

On 8/27/05, evilninja@gmx.net <evilninja@gmx.net> wrote:
> Ronald Moesbergen schrieb:
> > Hi,
> >
> > While running a vmware virtual machine from a reiser4 filesystem, my
> > machine consistently locks up and produces the following OOPS:
> >
> > kernel: Oops: 0002 [#1]
> > kernel: PREEMPT SMP
> > kernel: last sysfs file: /class/vc/vcsa8/dev
> > kernel: Modules linked in: vmnet vmmon nvidia
> 
> is this reproducible and is it reproducible when you don't run this
> inside vmware? the tainted kernel could make it hard to find a solution
> here...
> 
> Christian.
>

Re: Reiser4 oops with 2.6.13-rc6-mm2

[Ronald Moesbergen]

Well, I just successfully reproduced this without the nvidia module
loaded. The oops looks very similar:

 kernel: Unable to handle kernel NULL pointer dereference at virtual
address 00000000
 kernel:  printing eip:
 kernel: b01a1317
 kernel: *pde = 00000000
 kernel: Oops: 0002 [#1]
 kernel: PREEMPT SMP
 kernel: last sysfs file: /class/vc/vcsa8/dev
 kernel: Modules linked in: vmnet vmmon
 kernel: CPU:    1
 kernel: EIP:    0060:[lock_object+84/127]    Tainted: P      VLI
 kernel: EIP:    0060:[<b01a1317>]    Tainted: P      VLI
 kernel: EFLAGS: 00013202   (2.6.13-rc6-mm2)
 kernel: EIP is at lock_object+0x54/0x7f
 kernel: eax: de86be94   ebx: 00000000   ecx: ec287e80   edx: ec287e94
 kernel: esi: eae23d80   edi: ec287ec8   ebp: ec287ec8   esp: ec287b48
 kernel: ds: 007b   es: 007b   ss: 0068
 kernel: Process vmware-vmx (pid: 5580, threadinfo=ec286000 task=eecc9a30)
 kernel: Stack: eae23d80 00000000 ec287ec8 b01a16e7 ec287ec8 eae23db4
00000000 eae23df0
 kernel:        b01a17ba ec287ec8 00000000 00000000 00000001 00000000
00000000 ec287ec8
 kernel:        eae23d80 eae23df0 b01a18ad ec287ec8 00000000 ffffffff
ffffffff ec287c88
 kernel: Call Trace:
 kernel:  [lock_tail+80/113] lock_tail+0x50/0x71
 kernel:  [<b01a16e7>] lock_tail+0x50/0x71
 kernel:  [longterm_lock_tryfast+178/229] longterm_lock_tryfast+0xb2/0xe5
 kernel:  [<b01a17ba>] longterm_lock_tryfast+0xb2/0xe5
 kernel:  [longterm_lock_znode+192/695] longterm_lock_znode+0xc0/0x2b7
 kernel:  [<b01a18ad>] longterm_lock_znode+0xc0/0x2b7
 kernel:  [cbk_cache_scan_slots+320/754] cbk_cache_scan_slots+0x140/0x2f2
 kernel:  [<b01b228d>] cbk_cache_scan_slots+0x140/0x2f2
 kernel:  [schedule+1320/3501] schedule+0x528/0xdad
 kernel:  [<b039568c>] schedule+0x528/0xdad
 kernel:  [cbk_cache_search+39/92] cbk_cache_search+0x27/0x5c
 kernel:  [<b01b2466>] cbk_cache_search+0x27/0x5c
 kernel:  [coord_by_handle+16/41] coord_by_handle+0x10/0x29
 kernel:  [<b01b110f>] coord_by_handle+0x10/0x29
 kernel:  [object_lookup+188/230] object_lookup+0xbc/0xe6
 kernel:  [<b01b10d5>] object_lookup+0xbc/0xe6
 kernel:  [find_file_item+322/482] find_file_item+0x142/0x1e2
 kernel:  [<b01e4d71>] find_file_item+0x142/0x1e2
 kernel:  [read_file+185/530] read_file+0xb9/0x212
 kernel:  [<b01e6d83>] read_file+0xb9/0x212
 kernel:  [txn_begin+29/33] txn_begin+0x1d/0x21
 kernel:  [<b01a74b1>] txn_begin+0x1d/0x21
 kernel:  [read_unix_file+532/831] read_unix_file+0x214/0x33f
 kernel:  [<b01e70fe>] read_unix_file+0x214/0x33f
 kernel:  [init_context+117/189] init_context+0x75/0xbd
 kernel:  [<b01a47dd>] init_context+0x75/0xbd
 kernel:  [reiser4_read+143/255] reiser4_read+0x8f/0xff
 kernel:  [<b01baade>] reiser4_read+0x8f/0xff
 kernel:  [vfs_read+328/333] vfs_read+0x148/0x14d
 kernel:  [<b015e7e6>] vfs_read+0x148/0x14d
 kernel:  [sys_pread64+112/125] sys_pread64+0x70/0x7d
 kernel:  [<b015ebd4>] sys_pread64+0x70/0x7d
 kernel:  [sysenter_past_esp+84/117] sysenter_past_esp+0x54/0x75
 kernel:  [<b0102d9b>] sysenter_past_esp+0x54/0x75
 kernel: Code: 04 89 71 08 8d 51 0c 8b 58 04 89 41 0c 89 50 04 89 13
89 5a 04 8d 51 14 83 47 18 01 8b 86 84 00 00 00 8b 58 04 89 41 14 89
50 04 <89> 13 89 5a 04 c7 01 00 00 00 00 8b 57 0c 85 d2 74 04 83 46 7c
 kernel:  <3>BUG: vmware-vmx[5580] exited with nonzero preempt_count 1!

One more thing that might be of interest: the reiser4 partition is on
an md-raid0 volume. Hope this helps.

Ronald.


On 8/27/05, evilninja@gmx.net <evilninja@gmx.net> wrote:
> Ronald Moesbergen schrieb:
> > Hi,
> >
> > While running a vmware virtual machine from a reiser4 filesystem, my
> > machine consistently locks up and produces the following OOPS:
> >
> > kernel: Oops: 0002 [#1]
> > kernel: PREEMPT SMP
> > kernel: last sysfs file: /class/vc/vcsa8/dev
> > kernel: Modules linked in: vmnet vmmon nvidia
> 
> is this reproducible and is it reproducible when you don't run this
> inside vmware? the tainted kernel could make it hard to find a solution
> here...
> 
> Christian.
>

Re: Reiser4 oops with 2.6.13-rc6-mm2

[Vladimir V. Saveliev]

Hello

Ronald Moesbergen wrote:
> Hi,
> 
> Thanks for responding. To clarify: I'm not running linux inside
> vmware, I'm running vmware on linux and the vmware images are on a
> reiser4 partition. I'm running windows inside vmware. I can certainly
> reproduce this,

Yes, vmware + reiser4 problem is close to top of things to be fixed.

 but haven't tried without the nvidia module. I will do
> that right now and report back.
> 
> Thanks,
> Ronald.
> 
> On 8/27/05, evilninja@gmx.net <evilninja@gmx.net> wrote:
>>Ronald Moesbergen schrieb:
>>>Hi,
>>>
>>>While running a vmware virtual machine from a reiser4 filesystem, my
>>>machine consistently locks up and produces the following OOPS:
>>>
>>>kernel: Oops: 0002 [#1]
>>>kernel: PREEMPT SMP
>>>kernel: last sysfs file: /class/vc/vcsa8/dev
>>>kernel: Modules linked in: vmnet vmmon nvidia
>>is this reproducible and is it reproducible when you don't run this
>>inside vmware? the tainted kernel could make it hard to find a solution
>>here...
>>
>>Christian.
>>
> 
> 

Re: Reiser4 oops with 2.6.13-rc6-mm2

[evilninja]

Ronald Moesbergen schrieb:
> Well, I just successfully reproduced this without the nvidia module
> loaded. The oops looks very similar:
[...]
>  kernel: Modules linked in: vmnet vmmon
>  kernel: CPU:    1
>  kernel: EIP:    0060:[lock_object+84/127]    Tainted: P      VLI

well, the kernel is still tainted - by the vmware modules. but as Vladimir
pointed out, there seem to be vmware related issues in reiser4. so you
could wait for the fix or try to reproduce with an untainted kernel (maybe
it's a completly different problem which needs to be fixed).

Christian.
-- 
BOFH excuse #432:

Borg nanites have infested the server

Re: Reiser4 oops with 2.6.13-rc6-mm2

[Ronald Moesbergen]

OK, but the tainting is because the license is different from GPL, not
because of lack of source code (vmware modules source code is
available). Anyway, I'll wait for a fix. Let me know if there's
anything I can do to help.

Ronald.

On 8/28/05, evilninja <evilninja@gmx.net> wrote:
> Ronald Moesbergen schrieb:
> > Well, I just successfully reproduced this without the nvidia module
> > loaded. The oops looks very similar:
> [...]
> >  kernel: Modules linked in: vmnet vmmon
> >  kernel: CPU:    1
> >  kernel: EIP:    0060:[lock_object+84/127]    Tainted: P      VLI
> 
> well, the kernel is still tainted - by the vmware modules. but as Vladimir
> pointed out, there seem to be vmware related issues in reiser4. so you
> could wait for the fix or try to reproduce with an untainted kernel (maybe
> it's a completly different problem which needs to be fixed).
> 
> Christian.
> --
> BOFH excuse #432:
> 
> Borg nanites have infested the server
>

Re: Reiser4 oops with 2.6.13-rc6-mm2

[michael chang]

I know this probably doesn't fix your problem, but there's a program
called QEMU that might perform similar functionality to VMWare; does
it have the same issues?  Maybe you want to try that, and see if you
have the same problem or not...

On 8/28/05, Ronald Moesbergen <ronald.moesbergen@gmail.com> wrote:
> OK, but the tainting is because the license is different from GPL, not
> because of lack of source code (vmware modules source code is
> available). Anyway, I'll wait for a fix. Let me know if there's
> anything I can do to help.
> 
> Ronald.
> 
> On 8/28/05, evilninja <evilninja@gmx.net> wrote:
> > Ronald Moesbergen schrieb:
> > > Well, I just successfully reproduced this without the nvidia module
> > > loaded. The oops looks very similar:
> > [...]
> > >  kernel: Modules linked in: vmnet vmmon
> > >  kernel: CPU:    1
> > >  kernel: EIP:    0060:[lock_object+84/127]    Tainted: P      VLI
> >
> > well, the kernel is still tainted - by the vmware modules. but as Vladimir
> > pointed out, there seem to be vmware related issues in reiser4. so you
> > could wait for the fix or try to reproduce with an untainted kernel (maybe
> > it's a completly different problem which needs to be fixed).
> >
> > Christian.
> > --
> > BOFH excuse #432:
> >
> > Borg nanites have infested the server
> >
> 


-- 
~Mike
 - Just my two cents
 - No man is an island, and no man is unable.

Re: Reiser4 oops with 2.6.13-rc6-mm2

[David Masover]

michael chang wrote:
> I know this probably doesn't fix your problem, but there's a program
> called QEMU that might perform similar functionality to VMWare; does
> it have the same issues?

Sort of.  And more.

I don't know how fast VMWare is, but qemu is pretty slow, as it does
actual emulation.  There is an addon called "kqemu", which is not GPL
and has no source available, which does stuff to the host kernel to
allow qemu to approach native speeds (by avoiding emulation), but that
seems to be just the CPU -- other devices (disk, network, and so on) are
still emulated.

And, even with the proprietary addon, qemu is probably slower than VMWare.