[Unionfs] SMP machines+acpi = segfaults
Tomas M
tomas at linux-live.org
Mon Mar 6 08:33:03 EST 2006
I compiled kernel with frame pointers, like Jan E. suggested,
and the stack trace seems different now.
Using latest unionfs CVS (yesterday) and kernel 2.6.16pre5
Seems to me like the bug is NOT in squashfs, because both segfaults I
logged (occurs at the same time, maybe on both processors??) seems to
refer to unionfs_d_revalidate. Then the log stops (and the system freezes).
Note: this doesn't occur on single processor machine.
Please see attachment with the log
As far as I can see the segfaults I'm experiencing from time to time,
for the last one year, it's almost every time related to
unionfs_d_revalidate.
Is that function hard to implement correctly? Shouldn't it be rewriten?
Thank you for your consideration.
Tomas
Jan Engelhardt wrote:
>> Hello,
>> would you please tell me in which section is this option? Thank you very much
>>
>>
>
> Kernel hacking > Kernel debugging > Compile the kernel with frame pointers
> (CONFIG_FRAME_POINTER in .config)
>
>
>> Tomas
>>
>>
>> Jan Engelhardt wrote:
>>
>>>> Hello,
>>>>
>>>>
>>>>> If this is the case, then there could be a bug in Squashfs 3.0.
>>>>>
>>>>>
>>>> It can also be caused by unionfs, which could be unable to handle some
>>>> squashfs
>>>> 3 feature properly.
>>>> I saw this already in the past, sometimes unionfs needs some special
>>>> handling
>>>> for special filesystems,
>>>> for example for NFS.
>>>>
>>>>
>>> Enable the "compile with frame pointers" thing in the kernel, which might
>>> could 'fix' the impossible stack trace.
>>>
>>>
>>> Jan Engelhardt
>>>
>>>
>
> Jan Engelhardt
>
-------------- next part --------------
Mar 6 14:17:38 (none) kernel: Unable to handle kernel NULL pointer dereference at virtual address 00000079
Mar 6 14:17:38 (none) kernel: printing eip:
Mar 6 14:17:38 (none) kernel: f8ceeae8
Mar 6 14:17:38 (none) kernel: *pde = 00000000
Mar 6 14:17:38 (none) kernel: Oops: 0000 [#1]
Mar 6 14:17:38 (none) kernel: SMP
Mar 6 14:17:38 (none) kernel: Modules linked in: nls_iso8859_2 nls_iso8859_1 nls_cp437 unionfs squashfs
Mar 6 14:17:38 (none) kernel: CPU: 1
Mar 6 14:17:38 (none) kernel: EIP: 0060:[<f8ceeae8>] Not tainted VLI
Mar 6 14:17:38 (none) kernel: EFLAGS: 00010206 (2.6.16-rc5 #2)
Mar 6 14:17:38 (none) kernel: EIP is at unionfs_d_revalidate+0x2b8/0x440 [unionfs]
Mar 6 14:17:38 (none) kernel: eax: 0000002d ebx: 00000000 ecx: f7ddce00 edx: 0000000a
Mar 6 14:17:38 (none) kernel: esi: f6ec02f8 edi: f711c114 ebp: c21b9e14 esp: c21b9de4
Mar 6 14:17:38 (none) kernel: ds: 007b es: 007b ss: 0068
Mar 6 14:17:38 (none) kernel: Process xcompmgr (pid: 4555, threadinfo=c21b8000 task=c23e3030)
Mar 6 14:17:38 (none) kernel: Stack: <0>f6559070 c0175a40 f711c114 00000000 00000000 00000001 00000008 00000001
Mar 6 14:17:38 (none) kernel: c21b9f44 f711c114 c21b9e68 c21b9f44 c21b9e20 f8ceec89 f711c114 c21b9e3c
Mar 6 14:17:38 (none) kernel: c016b66c c2135d40 c21b9e70 f6d31006 00000004 c21b9e70 c21b9e8c c016b79f
Mar 6 14:17:38 (none) kernel: Call Trace:
Mar 6 14:17:38 (none) kernel: [<c01040d8>] show_stack_log_lvl+0x98/0xc0
Mar 6 14:17:38 (none) kernel: [<c0104267>] show_registers+0x137/0x1b0
Mar 6 14:17:38 (none) kernel: [<c0104463>] die+0xf3/0x190
Mar 6 14:17:38 (none) kernel: [<c011746c>] do_page_fault+0x36c/0x598
Mar 6 14:17:38 (none) kernel: [<c0103d1f>] error_code+0x4f/0x60
Mar 6 14:17:38 (none) kernel: [<f8ceec89>] unionfs_d_revalidate_wrap+0x19/0x30 [unionfs]
Mar 6 14:17:38 (none) kernel: [<c016b66c>] do_lookup+0x5c/0x90
Mar 6 14:17:38 (none) kernel: [<c016b79f>] __link_path_walk+0xff/0xcd0
Mar 6 14:17:38 (none) kernel: [<c016c3b2>] link_path_walk+0x42/0xd0
Mar 6 14:17:38 (none) kernel: [<c016c75d>] do_path_lookup+0x11d/0x260
Mar 6 14:17:38 (none) kernel: [<c016cb60>] __user_walk_fd+0x40/0x60
Mar 6 14:17:38 (none) kernel: [<c015c677>] sys_faccessat+0x87/0x150
Mar 6 14:17:38 (none) kernel: [<c015c752>] sys_access+0x12/0x20
Mar 6 14:17:38 (none) kernel: [<c0102c29>] syscall_call+0x7/0xb
Mar 6 14:17:38 (none) kernel: Code: 83 c4 24 5b 5e 5f 5d c3 8b 55 f0 ff d1 85 c0 75 07 c7 45 ec 00 00 00 00 43 3b 5d e8 7f 20 8b 77 54 8b 46 28 8b 04 98 85 c0 74 ed <8b> 50 4c 85 d2 74 e6 8b 0a 85 c9 75 d0 43 3b 5d e8 7e e0 8b 7f
Mar 6 14:17:38 (none) kernel: <1>Unable to handle kernel NULL pointer dereference at virtual address 0000005e
Mar 6 14:17:38 (none) kernel: printing eip:
Mar 6 14:17:38 (none) kernel: f8ceeae8
Mar 6 14:17:38 (none) kernel: *pde = 00000000
Mar 6 14:17:38 (none) kernel: Oops: 0000 [#2]
Mar 6 14:17:38 (none) kernel: SMP
Mar 6 14:17:38 (none) kernel: Modules linked in: nls_iso8859_2 nls_iso8859_1 nls_cp437 unionfs squashfs
Mar 6 14:17:38 (none) kernel: CPU: 0
Mar 6 14:17:39 (none) kernel: EIP: 0060:[<f8ceeae8>] Not tainted VLI
Mar 6 14:17:39 (none) kernel: EFLAGS: 00010206 (2.6.16-rc5 #2)
Mar 6 14:17:39 (none) kernel: EIP is at unionfs_d_revalidate+0x2b8/0x440 [unionfs]
Mar 6 14:17:39 (none) kernel: eax: 00000012 ebx: 00000002 ecx: f7ddce00 edx: 0000000a
Mar 6 14:17:39 (none) kernel: esi: f6ec0350 edi: f7ab9094 ebp: f6237dd0 esp: f6237da0
Mar 6 14:17:39 (none) kernel: ds: 007b es: 007b ss: 0068
Mar 6 14:17:39 (none) kernel: Process startkde (pid: 4529, threadinfo=f6236000 task=f7c40030)
Mar 6 14:17:39 (none) kernel: Stack: <0>f65593c0 c0175a40 f7ab9094 00000002 00000000 00000001 00000004 00000001
Mar 6 14:17:39 (none) kernel: f6237ef4 f7ab9094 f6237e24 f6237ef4 f6237ddc f8ceec89 f7ab9094 f6237df8
Mar 6 14:17:39 (none) kernel: c016b66c c2135d40 f6237e2c f6879005 00000003 f6237e2c f6237e48 c016b79f
Mar 6 14:17:39 (none) kernel: Call Trace:
Mar 6 14:17:39 (none) kernel: [<c01040d8>] show_stack_log_lvl+0x98/0xc0
Mar 6 14:17:39 (none) kernel: [<c0104267>] show_registers+0x137/0x1b0
Mar 6 14:17:39 (none) kernel: [<c0104463>] die+0xf3/0x190
Mar 6 14:17:39 (none) kernel: [<c011746c>] do_page_fault+0x36c/0x598
Mar 6 14:17:39 (none) kernel: [<c0103d1f>] error_code+0x4f/0x60
Mar 6 14:17:39 (none) kernel: [<f8ceec89>] unionfs_d_revalidate_wrap+0x19/0x30 [unionfs]
Mar 6 14:17:39 (none) kernel: [<c016b66c>] do_lookup+0x5c/0x90
Mar 6 14:17:39 (none) kernel: [<c016b79f>] __link_path_walk+0xff/0xcd0
Mar 6 14:17:39 (none) kernel: [<c016c3b2>] link_path_walk+0x42/0xd0
Mar 6 14:17:39 (none) kernel: [<c016c75d>] do_path_lookup+0x11d/0x260
Mar 6 14:17:39 (none) kernel: [<c016cb60>] __user_walk_fd+0x40/0x60
Mar 6 14:17:39 (none) kernel: [<c0167359>] vfs_stat_fd+0x19/0x50
Mar 6 14:17:39 (none) kernel: [<c01673a1>] vfs_stat+0x11/0x20
Mar 6 14:17:39 (none) kernel: [<c0167a24>] sys_stat64+0x14/0x30
Mar 6 14:17:39 (none) kernel: [<c0102c29>] syscall_call+0x7/0xb
Mar 6 14:17:39 (none) kernel: Code: 83 c4 24 5b 5e 5f 5d c3 8b 55 f0 ff d1 85 c0 75 07 c7 45 ec 00 00 00 00 43 3b 5d e8 7f 20 8b 77 54 8b 46 28 8b 04 98 85 c0 74 ed <8b> 50 4c 85 d2 74 e6 8b 0a 85 c9 75 d0 43 3b 5d e8 7e e0 8b 7f
More information about the unionfs
mailing list