[2.6.12-rc4] Oops in reiserfs_panic [please CC]

[2.6.12-rc4] Oops in reiserfs_panic [please CC]

[Fabio Rosciano]

Hi list,

I have tried to compile 2.6.12-rc4 on my iBook G4 running Debian Sid, 
with custom 2.6.9 kernel.
2.6.9 works absolutely wonderful on this machine, so I used the same 
.config file to compile 2.6.12-rc4, downloaded as a complete tarball.
Upon rebooting with the new kernel, my reiserfs partition (/ actually) 
is correctly seen and checked, but then (before syslogd starts) I get an 
Oops. I transcribe here the first lines:

REISERFS: panic
Kernel BUG in reiserfs_panic at fs/reiserfs/panic.c:362!
Oops: Exception in kernel mode, sig:5 [#1]

If you need any other info I will be glad to report them. Thanks for any 
help.

Re: [2.6.12-rc4] Oops in reiserfs_panic [please CC]

[Fabio Rosciano]

Vladimir Saveliev wrote:
> yes, please provide more information:
> does 2.6.9 still work corectly?

Absolutely wonderful, as it has ever done.

> does the oops come up every time  you boot with 2.6.12-rc4?

Yes.

> please ksymoops the oops output

I have never used this tool, but from the man page looks like it needs 
the syslogd stuff, and the oops occurs right before syslogd is started, 
so I have to copy it by hand :(
Ok let's go:

Checking internal tree... finished

REISERFS: panic (device NULL superblock): reiserfs [1795]: assertion !( 
comp_keys( &MAX_KEY, p_s_key ) && ! key_in_buffer(p_s_search_path, 
p_s_key, p_s_sb) ) failed at fs/reiserfs/stree.c:685:search_by_key: 
PAP-5130: key is not in the buffer

kernel BUG in reiserfs_panic at fs/reiserfs/prints.c:362!
Oops: Exception in kernel mode, sig: 5 [#1]
PREEMPT
NIP: C00C6C18 LR: C00C6C18 SP:E7BD7B80 REGS: e7bd7ad0 TRAP: 0700 Not tainted
MSR: 00029032 EE: 1 PR: 0 FP: 0 ME: 1 IR/DR: 11
TASK = e77f86f0[1795] 'mount' THREAD: e7bd6000
Last syscall: 21
GPR00: C00C6C18 E7BD7B80 E77F86F0 000000FB 00002798 FFFFFFFF E7C6B4D0 
00000008
GPR08: 0000000B C04D0000 00000008 E7BD6000 88000424 1002A544 00000000 
E7BD7D28
GPR16: 100241F8 00000004 00000001 E7BD7D60 00000000 E7BD7D10 E7D6A200 
10024218
GPR24: 7F927F3E E7D6A200 E7BD7D60 E7BD7D60 E7BD7D60 C0371850 E7BD7D10 
00000000
NIP [c00c6c18] reiserfs_panic+0x70/0x9c
LR [c00c6c18] reiserfs_panic+0x70/0x9c
Call Trace:
[c00d3f98] search_by_key+0x2898/0x3200
[c00c2b84] finish_unfinished+0x70/0x428
[c00c4128] reiserfs_remount+0x2b8/0x2e0
[c006c6c8] do_remount_sb+0xbc/0x14c
[c0087398] do_remount+oxac/0x108
[c0087fe4] do_mount+0x184/0x190
[c0088474] sys_mount+0xa8/0xfc
[c0004680] ret_from_syscall+0x0/0xfc
/etc/rcS.d/S10checkroot.sh: line 290: 1795 trace/breakpoint trap  mount 
-n -o remount,$rootopts,$rootmode $fstabroot / 2>/dev/null

> send all output which might be related to this oops

This is my partition table:

/dev/hda
         #                    type name                 length   base 
   ( size )  system
/dev/hda1     Apple_partition_map Apple                    63 @ 1 
  ( 31.5k)  Partition map
/dev/hda2         Apple_Bootstrap untitled               2048 @ 64 
  (  1.0M)  NewWorld bootblock
/dev/hda3         Apple_Bootstrap /                  10240000 @ 2112 
  (  4.9G)  NewWorld bootblock
/dev/hda4               Apple_HFS Apple_HFS_Untitled_3 20709296 @ 
57430848 (  9.9G)  HFS
/dev/hda5         Apple_UNIX_SVR2 swap                 524288 @ 10242112 
(256.0M)  Linux swap
/dev/hda6         Apple_UNIX_SVR2 untitled           46664448 @ 10766400 
( 22.3G)  Linux native
/dev/hda7              Apple_Free Extra                    16 @ 78140144 
(  8.0k)  Free space

Block size=512, Number of Blocks=78140160
DeviceType=0x0, DeviceId=0x0

This is the relevant kernel config:

CONFIG_REISERFS_FS=y
CONFIG_REISERFS_CHECK=y
CONFIG_REISERFS_PROC_INFO=y
CONFIG_REISERFS_FS_XATTR=y
CONFIG_REISERFS_FS_POSIX_ACL=y
CONFIG_REISERFS_FS_SECURITY=y

And the checkroot.sh script works good with 2.6.9, so I don't think this 
is related.

Hope this helps, please ask if you need anything else! Thank you!

-- 
Best Regards, Jack
Debian Sid PPC

Re: [2.6.12-rc4] Oops in reiserfs_panic [please CC]

[Jesse D Zbikowski]

This looks to be the same bug reported by two Gentoo users against
2.6.12-rc4, as a regression from -rc3.  One reported data corruption.
You should be able to work around by compiling with

CONFIG_REISERFS_CHECK=n

http://bugs.gentoo.org/show_bug.cgi?id=91968

Here is the oops:

ReiserFS: hda7: Using r5 hash to sort names
REISERFS: panic (device Null superblock): reiserfs[4540]: assertion !(
comp_keys( &MAX_KEY, p_s_key ) && ! key_in_buffer(p_s_search_path, p_s_key,
p_s_sb) ) failed at fs/reiserfs/stree.c:685:search_by_key: PAP-5130: key is not
in the buffer

------------[ cut here ]------------
kernel BUG at fs/reiserfs/prints.c:362!
invalid operand: 0000 [#1]
PREEMPT
Modules linked in: eeprom i2c_dev asb100 i2c_sensor i2c_core unix
CPU:    0
EIP:    0060:[<c01b05fe>]    Not tainted VLI
EFLAGS: 00010282   (2.6.12-rc4)
EIP is at reiserfs_panic+0x51/0x76
eax: 000000fb   ebx: c0405473   ecx: 000029f2   edx: c04b6701
esi: 00000000   edi: 00000140   ebp: ef2ebd78   esp: ef2ebbcc
ds: 007b   es: 007b   ss: 0068
Process mount (pid: 4540, threadinfo=ef2ea000 task=ef2540a0)
Stack: c04108b4 c0405473 c054ffa0 ffffffff ffffffff c03f008d c01b9daa 00000000
       c0415454 000011bc 000002ad c03f010c 00000000 00000001 ef2ebd7c c01b8933
       eeba0000 ffffffff 00000000 ef2ea000 ef2ea000 ef2ea000 ef2ea000 ef2ea000
Call Trace:
 [<c01b9daa>] search_by_key+0x13cd/0x1bc4
 [<c01b8933>] is_tree_node+0x6c/0x71
 [<c013ecf8>] __alloc_pages+0x173/0x3d8
 [<c011a0a6>] call_console_drivers+0x67/0x13b
 [<c01ad069>] finish_unfinished+0x9f/0x3a2
 [<c012e1a5>] autoremove_wake_function+0x0/0x57
 [<c01c4e80>] do_journal_end+0x7f3/0x966
 [<c025d3c3>] vsprintf+0x27/0x2b
 [<c01af7cc>] reiserfs_fill_super+0x6c6/0x77e
 [<c01a14c9>] reiserfs_init_locked_inode+0x0/0x16
 [<c015f82e>] sb_set_blocksize+0x2e/0x5e
 [<c015f1f8>] get_sb_bdev+0xe0/0x145
 [<c01742dc>] alloc_vfsmnt+0x9c/0xd1
 [<c01af8f0>] get_super_block+0x2f/0x33
 [<c01af106>] reiserfs_fill_super+0x0/0x77e
 [<c015f46c>] do_kern_mount+0x63/0xe9
 [<c0175437>] do_new_mount+0x9e/0xf7
 [<c0175b27>] do_mount+0x19d/0x1bb
 [<c0175933>] copy_mount_options+0x60/0xb7
 [<c0175ef4>] sys_mount+0x9f/0xd7
 [<c0102a27>] sysenter_past_esp+0x54/0x75
Code: 24 8d be 40 01 00 00 e8 e2 fc ff ff 85 f6 89 d8 c7 44 24 08 a0 ff 54 c0 c7
04 24 b4 08 41 c0 0f 45 c7 89 44 24 04 e8 42 9c f6 ff <0f> 0b 6a 01 36 5a 40 c0
85 f6 c7 44 24 08 a0 ff 54 c0 c7 04 24

Re: [2.6.12-rc4] Oops in reiserfs_panic [please CC]

[Fabio Rosciano]

Jesse D Zbikowski wrote:
> This looks to be the same bug reported by two Gentoo users against
> 2.6.12-rc4, as a regression from -rc3.  One reported data corruption.
> You should be able to work around by compiling with
> 
> CONFIG_REISERFS_CHECK=n
> 
> http://bugs.gentoo.org/show_bug.cgi?id=91968

Yep, it worked. Nasty bug though.

Thanks, for any other info just ask!

-- 
Best Regards, Jack
Debian Sid PPC