lkml.org 
[lkml]   [2008]   [Apr]   [3]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [BUG] 2.6.24.4 kernel bug while running ftest03
Nick Piggin wrote:
> On Thu, Apr 03, 2008 at 02:53:44PM +0200, Jan Kara wrote:
>> Hello,
>>
>> I've just found that Nick has been recently fixing this function,
>> Nick?
>
> Hmm, yes it looks like an earlier oops but that should be fixed in
> 2.6.24.4... I can't see a way it could advance off the end of the
> iovec now. Sigh.
>
> Is the problem easy to reproduce (preferably without the gov patch
> but even with the patch...).
>
> I couldn't reproduce the bug with ftest03.

I tried without the gcov patch and I am not able to reproduce it,
>
>>
>> Honza
>>
>>>> The kernel panic's, while running the LTP ftest03 on the 2.6.24.4 kernel
>>>> compiled with the gcov patches from LTP.
>>>> (http://ltp.cvs.sourceforge.net/ltp/utils/analysis/gcov-kernel/linux-2.6.24-gcov.patch?view=markup)
>>> Are you able to reproduce the problem without this patch? The patch is
>>> nontrivial...
>>>
>>>> BUG: unable to handle kernel paging request at virtual address f8000004
>>> The address is a bit strange. Looks like we have advanced iov beyond
>>> the area we have allocated. Can you send me disassembly of the function
>>> iov_iter_advance() of your kernel? Thanks.

Sorry, I do not have the kernel, i will try reproducing the get it.
>>>
>>>> printing eip: c01c2836 *pdpt = 0000000000004001 *pde = 0000000000000000
>>>> Oops: 0000 [#1] SMP
>>>>
>>>> Modules linked in: hidp rfcomm l2cap bluetooth dm_multipath nvram via686a hwmon i2c_viapro pcspkr mptsas scsi_transport_sas floppy tg3 parport_pc ac battery button dm_mirror dm_mod joydev sunrpc i2c_dev i2c_core autofs4 lp parport ipv6 md5 sg mptspi mptscsih mptbase scsi_transport_spi sd_mod scsi_mod ehci_hcd ohci_hcd uhci_hcdPid: 24160, comm: ftest03 Not tainted (2.6.24.4-gcov-autokern1 #1)
>>>>
>>>> EIP: 0060:[<c01c2836>] EFLAGS: 00010246 CPU: 1
>>>> EIP is at iov_iter_advance+0xe9/0x155
>>>> EAX: 00000000 EBX: f3457d98 ECX: 00000080 EDX: 00000000
>>>> ESI: 00000000 EDI: f8000000 EBP: f33f6d30 ESP: f3457d24
>>>> DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
>>>> Process ftest03 (pid: 24160, ti=f3457000 task=f6544590 task.ti=f3457000)
>>>> Stack: 00000800 0006b800 00000000 c01c5e79 0006b800 00000000 00000800 00000800
>>>> c7209030 c015cdea f3457e8c f7fb9180 f33f6d30 c07b1600 f33f6c68 00000000
>>>> 00000800 00000800 00000000 c07b1600 00000000 0000006b f3457da8 c02f11cd
>>>> Call Trace:
>>>> [<c01c5e79>] generic_file_buffered_write+0x29f/0xc05
>>>> [<c015cdea>] current_fs_time+0x4e/0x6f
>>>> [<c02f11cd>] ext3_readpages+0x0/0x35
>>>> [<c02f11cd>] ext3_readpages+0x0/0x35
>>>> [<c01d1da4>] __do_page_cache_readahead+0x268/0x386
>>>> [<c015cdea>] current_fs_time+0x4e/0x6f
>>>> [<c01c6f97>] __generic_file_aio_write_nolock+0x7b8/0x852
>>>> [<c01c70be>] generic_file_aio_write+0x8d/0x162
>>>> [<c01c7bfd>] generic_file_aio_read+0x292/0x31a
>>>> [<c02ecfc2>] ext3_file_write+0x32/0x17c
>>>> [<c020b830>] do_sync_readv_writev+0xcf/0x152
>>>> [<c01c5163>] wait_on_page_writeback_range+0x153/0x1e7
>>>> [<c017d0f7>] autoremove_wake_function+0x0/0x5f
>>>> [<c04f21b9>] copy_from_user+0xb1/0x135
>>>> [<c020b602>] rw_copy_check_uvector+0xd2/0x19f
>>>> [<c020c4f6>] do_readv_writev+0x152/0x328
>>>> [<c02ecf90>] ext3_file_write+0x0/0x17c
>>>> [<c079c11d>] _spin_unlock+0x13/0x22
>>>> [<c01b427e>] audit_syscall_entry+0x2eb/0x323
>>>> [<c020c758>] vfs_writev+0x8c/0x9e
>>>> [<c020cf6c>] sys_writev+0x8b/0x1a5
>>>> [<c0108a76>] syscall_call+0x7/0xb
>>>> =======================
>>>> Code: c0 00 29 c8 01 ce 89 43 0c 39 77 04 75 13 83 c7 08 83 05 98 63 a6 c0 01 83 15 9c 63 a6 c0 00 31 f6 29 ca 85 d2 0f 85 78 ff ff ff <83> 7f 04 00 75 29 83 05 a0 63 a6 c0 01 83 15 a4 63 a6 c0 00 83 EIP: [<c01c2836>] iov_iter_advance+0xe9/0x155 SS:ESP 0068:f3457d24---[ end trace af3ae442124c3d18 ]---


--
Thanks & Regards,
Kamalesh Babulal,
Linux Technology Center,
IBM, ISTL.


\
 
 \ /
  Last update: 2008-04-03 19:03    [W:0.055 / U:0.212 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site