Segfault mounting a reiserfs 3.6 partition

Segfault mounting a reiserfs 3.6 partition

[Christophe Greisberger]

[-- Attachment #1: Type: text/plain, Size: 4207 bytes --]

Hi,

I have a strange problem with a partition w/ reiserfs.
After a power loss, the partition has been damaged in a way reiserfs doesn't 
seem to handle correctly. When mounting the partition, there is a segfault.
The original system was a SuSE 8.0 (kernel 2.4.20)
I did a 'reiserfsck --check ' , then a 'reiserfsck --fix-fixable' : no 
errors found/fixed.
I then tried 'reiserfsck --rebuild-sb' and 'reiserfsck --rebuild-tree' : 
some errors fixed, still segfaulting.
Finally, 'reiserfsck --clean-attributes' : no errors found, still 
segfaulting.

I then tried with a recent SuSE live CD (10.0 - kernel 2.6.13) : same error.

Finally, I used sysresccd (kernel 2.6.16.10)  to save & restore the 
partition with partimage, hoping that it would rebuild a clean partition.
-The save was ok.
-I then reformatted the partition with r/w check - ok.
-The restoration worked.
I then tried to mount the partition : still the same error. Damn!

Here the error :
----------8<----------
# mount /dev/sda4 /mnt/temp1
ReiserFS: sda4: found reiserfs format "3.6" with standard journal
ReiserFS: sda4: using ordered data mode
ReiserFS: sda4: journal params: device sda4, size 8192, journal first block 
18, max trans len 1024,
max batch 900, max commit age 30, max trans age 30
ReiserFS: sda4: checking transaction log (sda4)
ReiserFS: sda4: Using r5 hash to sort names
------------[ cut here ]------------
kernel BUG at fs/reiserfs/journal.c:3168!
invalid opcode: 0000 [#1]
Modules linked in: i2c_nforce2 yellowfin via_velocity typhoon tg3 sungem_phy 
sundance starfire slip
slhc skge sis900 sis190 shaper s2io plip pcnet32 ns83820 ne3210 ne2k_pci 
natsemi hp100 hamachi fealnx epic100 eepro100 dl2k dgrs depca cassini bnx2 
b44 amd8111e 8390 8139cp 3c59x 3c509
CPU:    0
EIP:    0060:[<c01a5c4d>]    Not tainted VLI
EFLAGS: 00010246   (2.6.16.10-fd24 #1)
EIP is at journal_mark_dirty+0x215/0x240
eax: cfb5dc00   ebx: cfb5dc00   ecx: 00000002   edx: 00000000
esi: d0061600   edi: cf073e08   ebp: e0df5000   esp: cf09ddc4
ds: 007b   es: 007b   ss: 0068
Process mount (pid: 4720, threadinfo=cf09c000 task=dfae4570)
Stack: <0>00000002 00000000 cfb5dc00 ce7f5dac d0061600 d0061600 c019779c 
cf09de1c
       d0061600 cf073e08 00000000 cef17000 00000000 00000002 00000001 
00000000
       00000000 00000000 00000004 dfb875f0 cf09de74 c0367a02 d0061600 
00000001
Call Trace:
 [<c019779c>] reiserfs_fill_super+0x58c/0x664
 [<c0367a02>] snprintf+0x16/0x1c
 [<c014d8d7>] get_sb_bdev+0xbb/0x114
 [<c036842e>] strncpy_from_user+0x2e/0x4c
 [<c016016c>] copy_mount_options+0x44/0xa0
 [<c016061c>] sys_mount+0x74/0xb0
 [<c0102ce9>] syscall_call+0x7/0xb
Code: 8b 45 30 50 68 fc bd 6c c0 56 e8 bb 2d ff ff 83 c4 0c b8 01 00 00 00 
eb b3 50 52 68 40 be 6c c0 8b 54 24 28 ff 32 e8 4f 2e ff ff <0f> 0b 60 0c 
72 cd 6b c0 e9 0d fe ff ff 52 68 7c be 6c c0 8b 44
 Badness in do_exit at kernel/exit.c:802
 [<c0115af1>] do_exit+0x34d/0x358
 [<c01035ac>] do_trap+0x0/0xd4
 [<c0103828>] do_invalid_op+0x0/0x88
 [<c01038a2>] do_invalid_op+0x7a/0x88
 [<c0110907>] __wake_up_common+0x27/0x4c
 [<c01a5c4d>] journal_mark_dirty+0x215/0x240
 [<c067b395>] schedule+0x2dd/0x5d0
 [<c067bbb7>] cond_resched+0x37/0x50
 [<c01a5592>] do_journal_begin_r+0x56/0x2b8
 [<c0102f13>] error_code+0x4f/0x54
 [<c01a5c4d>] journal_mark_dirty+0x215/0x240
 [<c019779c>] reiserfs_fill_super+0x58c/0x664
 [<c0367a02>] snprintf+0x16/0x1c
 [<c014d8d7>] get_sb_bdev+0xbb/0x114
 [<c01980aa>] get_super_block+0x1a/0x20
 [<c0197210>] reiserfs_fill_super+0x0/0x664
 [<c014da8c>] do_kern_mount+0x3c/0x98
 [<c015fe02>] do_new_mount+0x62/0xa8
 [<c0160357>] do_mount+0x18f/0x1bc
 [<c036842e>] strncpy_from_user+0x2e/0x4c
 [<c016016c>] copy_mount_options+0x44/0xa0
 [<c016061c>] sys_mount+0x74/0xb0
 [<c0102ce9>] syscall_call+0x7/0xb
Segmentation fault
----------8<----------

I also googled a while, but I only found a post (was in 2005 I think), with 
the same error, but to which nobody answered.

An idea someone?

-- 
Christophe Greisberger
Zenon-Media GmbH
Phone: +49 (7852) 9133 / 34
Fax  : +49 (7852) 9133 / 35
--
Linux doesn't dictate how I work,
I dictate how Linux works.

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 2243 bytes --]

Re: Segfault mounting a reiserfs 3.6 partition

[Edward Shishkin]

Hello.
would you please upgrade your kernel to the latest stable version (2.6.18)
and try to mount again.

Thanks,
Edward.
 

Christophe Greisberger wrote:

>Hi,
>
>I have a strange problem with a partition w/ reiserfs.
>After a power loss, the partition has been damaged in a way reiserfs doesn't 
>seem to handle correctly. When mounting the partition, there is a segfault.
>The original system was a SuSE 8.0 (kernel 2.4.20)
>I did a 'reiserfsck --check ' , then a 'reiserfsck --fix-fixable' : no 
>errors found/fixed.
>I then tried 'reiserfsck --rebuild-sb' and 'reiserfsck --rebuild-tree' : 
>some errors fixed, still segfaulting.
>Finally, 'reiserfsck --clean-attributes' : no errors found, still 
>segfaulting.
>
>I then tried with a recent SuSE live CD (10.0 - kernel 2.6.13) : same error.
>
>Finally, I used sysresccd (kernel 2.6.16.10)  to save & restore the 
>partition with partimage, hoping that it would rebuild a clean partition.
>-The save was ok.
>-I then reformatted the partition with r/w check - ok.
>-The restoration worked.
>I then tried to mount the partition : still the same error. Damn!
>
>Here the error :
>----------8<----------
># mount /dev/sda4 /mnt/temp1
>ReiserFS: sda4: found reiserfs format "3.6" with standard journal
>ReiserFS: sda4: using ordered data mode
>ReiserFS: sda4: journal params: device sda4, size 8192, journal first block 
>18, max trans len 1024,
>max batch 900, max commit age 30, max trans age 30
>ReiserFS: sda4: checking transaction log (sda4)
>ReiserFS: sda4: Using r5 hash to sort names
>------------[ cut here ]------------
>kernel BUG at fs/reiserfs/journal.c:3168!
>invalid opcode: 0000 [#1]
>Modules linked in: i2c_nforce2 yellowfin via_velocity typhoon tg3 sungem_phy 
>sundance starfire slip
>slhc skge sis900 sis190 shaper s2io plip pcnet32 ns83820 ne3210 ne2k_pci 
>natsemi hp100 hamachi fealnx epic100 eepro100 dl2k dgrs depca cassini bnx2 
>b44 amd8111e 8390 8139cp 3c59x 3c509
>CPU:    0
>EIP:    0060:[<c01a5c4d>]    Not tainted VLI
>EFLAGS: 00010246   (2.6.16.10-fd24 #1)
>EIP is at journal_mark_dirty+0x215/0x240
>eax: cfb5dc00   ebx: cfb5dc00   ecx: 00000002   edx: 00000000
>esi: d0061600   edi: cf073e08   ebp: e0df5000   esp: cf09ddc4
>ds: 007b   es: 007b   ss: 0068
>Process mount (pid: 4720, threadinfo=cf09c000 task=dfae4570)
>Stack: <0>00000002 00000000 cfb5dc00 ce7f5dac d0061600 d0061600 c019779c 
>cf09de1c
>       d0061600 cf073e08 00000000 cef17000 00000000 00000002 00000001 
>00000000
>       00000000 00000000 00000004 dfb875f0 cf09de74 c0367a02 d0061600 
>00000001
>Call Trace:
> [<c019779c>] reiserfs_fill_super+0x58c/0x664
> [<c0367a02>] snprintf+0x16/0x1c
> [<c014d8d7>] get_sb_bdev+0xbb/0x114
> [<c036842e>] strncpy_from_user+0x2e/0x4c
> [<c016016c>] copy_mount_options+0x44/0xa0
> [<c016061c>] sys_mount+0x74/0xb0
> [<c0102ce9>] syscall_call+0x7/0xb
>Code: 8b 45 30 50 68 fc bd 6c c0 56 e8 bb 2d ff ff 83 c4 0c b8 01 00 00 00 
>eb b3 50 52 68 40 be 6c c0 8b 54 24 28 ff 32 e8 4f 2e ff ff <0f> 0b 60 0c 
>72 cd 6b c0 e9 0d fe ff ff 52 68 7c be 6c c0 8b 44
> Badness in do_exit at kernel/exit.c:802
> [<c0115af1>] do_exit+0x34d/0x358
> [<c01035ac>] do_trap+0x0/0xd4
> [<c0103828>] do_invalid_op+0x0/0x88
> [<c01038a2>] do_invalid_op+0x7a/0x88
> [<c0110907>] __wake_up_common+0x27/0x4c
> [<c01a5c4d>] journal_mark_dirty+0x215/0x240
> [<c067b395>] schedule+0x2dd/0x5d0
> [<c067bbb7>] cond_resched+0x37/0x50
> [<c01a5592>] do_journal_begin_r+0x56/0x2b8
> [<c0102f13>] error_code+0x4f/0x54
> [<c01a5c4d>] journal_mark_dirty+0x215/0x240
> [<c019779c>] reiserfs_fill_super+0x58c/0x664
> [<c0367a02>] snprintf+0x16/0x1c
> [<c014d8d7>] get_sb_bdev+0xbb/0x114
> [<c01980aa>] get_super_block+0x1a/0x20
> [<c0197210>] reiserfs_fill_super+0x0/0x664
> [<c014da8c>] do_kern_mount+0x3c/0x98
> [<c015fe02>] do_new_mount+0x62/0xa8
> [<c0160357>] do_mount+0x18f/0x1bc
> [<c036842e>] strncpy_from_user+0x2e/0x4c
> [<c016016c>] copy_mount_options+0x44/0xa0
> [<c016061c>] sys_mount+0x74/0xb0
> [<c0102ce9>] syscall_call+0x7/0xb
>Segmentation fault
>----------8<----------
>
>I also googled a while, but I only found a post (was in 2005 I think), with 
>the same error, but to which nobody answered.
>
>An idea someone?
>
>  
>

Re: Segfault mounting a reiserfs 3.6 partition

[Christophe Greisberger]

[-- Attachment #1: Type: text/plain, Size: 253 bytes --]

Hmmm.
I'll search for a Live CD with the latest kernel.


On Thursday, 19 October 2006 11:45, Edward Shishkin wrote:
> Hello.
> would you please upgrade your kernel to the latest stable version
> (2.6.18) and try to mount again.
>
> Thanks,
> Edward.
>

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 2243 bytes --]

Re: Segfault mounting a reiserfs 3.6 partition

[Christophe Greisberger]

[-- Attachment #1: Type: text/plain, Size: 484 bytes --]

Yesssssssssss!

I could mount the partition with the latest Gentoo live CD.
I'll now tar the data and reformat the partition.
Thanks for your help!

On Thursday, 19 October 2006 18:30, Christophe Greisberger wrote:
> Hmmm.
> I'll search for a Live CD with the latest kernel.
>
> On Thursday, 19 October 2006 11:45, Edward Shishkin wrote:
> > Hello.
> > would you please upgrade your kernel to the latest stable version
> > (2.6.18) and try to mount again.
> >
> > Thanks,
> > Edward.

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 2243 bytes --]

Re: Segfault mounting a reiserfs 3.6 partition

[Alexander Zarochentsev]

Hi,

On 20 October 2006 12:29, Christophe Greisberger wrote:
> Yesssssssssss!
>
> I could mount the partition with the latest Gentoo live CD.
> I'll now tar the data and reformat the partition.

likely it was the transaction id overflow bug (fixed in 2.6.17) , the fs 
itself is not broken.

it should work with 2.6.17+, it even may work with the older kernels 
after mount r/w and clean unmount.

> Thanks for your help!
>
> On Thursday, 19 October 2006 18:30, Christophe Greisberger wrote:
> > Hmmm.
> > I'll search for a Live CD with the latest kernel.
> >
> > On Thursday, 19 October 2006 11:45, Edward Shishkin wrote:
> > > Hello.
> > > would you please upgrade your kernel to the latest stable version
> > > (2.6.18) and try to mount again.
> > >
> > > Thanks,
> > > Edward.

-- 
Alex.

Re: Segfault mounting a reiserfs 3.6 partition

[Christian Kujau]

On Thu, 19 Oct 2006, Edward Shishkin wrote:
> would you please upgrade your kernel to the latest stable version (2.6.18)
> and try to mount again.

While I'm happy that Christophe can access his data with the latest 
kernel, did you have any special in mind when you suggested to upgrade 
from 2.4.20 or 2.6.13 (apart from the fact that both kernels are quite 
old)?

Thanks,
Christian.
-- 
BOFH excuse #257:

That would be because the software doesn't work.

Re: Segfault mounting a reiserfs 3.6 partition

[Christophe Greisberger]

> likely it was the transaction id overflow bug (fixed in 2.6.17) , the fs
> itself is not broken.
>
> it should work with 2.6.17+, it even may work with the older kernels
> after mount r/w and clean unmount.
Yes, the Live CD indeed has 2.6.17.

Christophe.