[Unionfs] repeatable kernel BUG in unionfs 2.1.3

Shaya Potter spotter at cs.columbia.edu
Tue Sep 11 18:55:18 EDT 2007


the sync ; sync ; sync seems to have helped, as I was able to create a 
union of 400+ branches w/o an issue, and performance seems reasonable.

Shaya Potter wrote:
> what I'm basically doing is
> setup union
> while [ 1 ]
>     chroot union apt-get install package_$i
>     remount add new top branch
>     remount mode->ro old top branch
> 
> I'm rerunning it w/ a "sync; sync; sync" b/w the chroot and the first 
> remount and haven't hit the BUG() yet.
> 
> Shaya Potter wrote:
>> Just compiled unionfs into 2.6.22.6 and have a repeatable BUG() in a 
>> stress test of mine that unionfs seems to tickle.
>>
>> My unionfs is slightly modified to allow up to 512 branches (simple 
>> header define change).  In this case, I was up to 85 branches when 
>> this occured.
>>
>> here's the backtrace
>>
>> ------------[ cut here ]------------
>> Kernel BUG at c0179fea [verbose debug info unavailable]
>> invalid opcode: 0000 [#1]
>> SMP
>> Modules linked in: nfsd exportfs lockd sunrpc af_packet dm_snapshot 
>> dm_mirror dm_mod sg sr_mod sd_mod scsi_mod ide_generic ide_cd cdrom 
>> i2c_piix4 psmouse i2c_core rtc ext3 jbd mbcache ide_disk serverworks 
>> tg3 ide_core unix
>> CPU:    0
>> EIP:    0060:[<c0179fea>]    Not tainted VLI
>> EFLAGS: 00010202   (2.6.22.6 #2)
>> EIP is at __block_write_full_page+0x179/0x358
>> eax: 8000183d   ebx: eb45a208   ecx: 00000014   edx: c1d7c940
>> esi: dfcebf60   edi: 00000000   ebp: ebbe748c   esp: dfcebd38
>> ds: 007b   es: 007b   fs: 00d8  gs: 0000  ss: 0068
>> Process pdflush (pid: 81, ti=dfcea000 task=dfc84a50 task.ti=dfcea000)
>> Stack: dfcebd5c c011542e 00000000 00000000 00000000 00001000 eb45a208 
>> 00000bff
>>        00000000 c014942e f88dd168 c1d7c940 00c00000 00000000 dfcebf60 
>> 00000000
>>        c017b7b7 dfcebf60 ebbe748c dfcebf60 f88dd168 c1d7c940 c1d7c940 
>> eb45a208
>> Call Trace:
>>  [<c011542e>] __wake_up+0x38/0x4e
>>  [<c014942e>] set_page_address+0xa2/0x15c
>>  [<f88dd168>] ext3_get_block+0x0/0xd7 [ext3]
>>  [<c017b7b7>] block_write_full_page+0xe3/0xee
>>  [<f88dd168>] ext3_get_block+0x0/0xd7 [ext3]
>>  [<f88ddb79>] ext3_ordered_writepage+0xb5/0x16c [ext3]
>>  [<f88ddaa4>] bget_one+0x0/0x7 [ext3]
>>  [<f88ddac4>] ext3_ordered_writepage+0x0/0x16c [ext3]
>>  [<c01a2765>] unionfs_writepage+0xd9/0x227
>>  [<c01445e7>] __writepage+0x8/0x21
>>  [<c01444e5>] write_cache_pages+0x1c5/0x2bf
>>  [<c01445df>] __writepage+0x0/0x21
>>  [<c014462b>] generic_writepages+0x2b/0x2d
>>  [<c014466a>] do_writepages+0x3d/0x43
>>  [<c017572e>] __sync_single_inode+0x55/0x1bd
>>  [<c01758dd>] __writeback_single_inode+0x47/0x1a5
>>  [<c0114888>] load_balance_newidle+0x62/0x152
>>  [<c0175bc5>] sync_sb_inodes+0x18a/0x268
>>  [<c0175d42>] writeback_inodes+0x9f/0xa8
>>  [<c01440e3>] background_writeout+0x67/0xa1
>>  [<c0144bd3>] __pdflush+0xcb/0x17d
>>  [<c0144c85>] pdflush+0x0/0x2c
>>  [<c0144c85>] pdflush+0x0/0x2c
>>  [<c0144cad>] pdflush+0x28/0x2c
>>  [<c014407c>] background_writeout+0x0/0xa1
>>  [<c012b2a6>] kthread+0x6b/0x6f
>>  [<c012b23b>] kthread+0x0/0x6f
>>  [<c010339b>] kernel_thread_helper+0x7/0x10
>>  =======================
>> Code: 5b 5e 5f 5d c3 8b 44 24 2c e8 d0 4d fc ff 8b 4c 24 44 83 41 10 
>> 01 eb e1 89 da b8 01 00 00 00 e8 8d 18 00 00 83 44 24 10 01 eb ad <0f> 
>> 0b eb fe 8b 74 24 44 8b 7e 04 85 ff 75 06 f6 46 24 01 75 37
>> EIP: [<c0179fea>] __block_write_full_page+0x179/0x358 SS:ESP 
>> 0068:dfcebd38
>> _______________________________________________
>> unionfs mailing list: http://unionfs.filesystems.org/
>> unionfs at mail.fsl.cs.sunysb.edu
>> http://www.fsl.cs.sunysb.edu/mailman/listinfo/unionfs
> _______________________________________________
> unionfs mailing list: http://unionfs.filesystems.org/
> unionfs at mail.fsl.cs.sunysb.edu
> http://www.fsl.cs.sunysb.edu/mailman/listinfo/unionfs


More information about the unionfs mailing list