I got this on 2.5.63 while running the "find" command:
kernel BUG at drivers/block/deadline-iosched.c:145!
invalid operand: 0000
CPU: 3
EIP: 0060:[<c026f8a1>] Not tainted
EFLAGS: 00010046
EIP is at deadline_find_drq_hash+0x41/0xb0
eax: f76fe060 ebx: f76fe058 ecx: f76fe060 edx: 00000000
esi: f76fe060 edi: f76fe044 ebp: f76fe050 esp: f7e27c88
ds: 007b es: 007b ss: 0068
Process pdflush (pid: 16, threadinfo=f7e26000 task=c3b39300)
Stack: 010d7ce6 00000000 f7fd2ae0 f7fd2ae0 f7fd2ae0 f7747f80 c026fbc9
f7747f80
010d7ce6 00000000 f7747f80 00000000 f7759000 f7759000 00000008
c026a438
f7759000 f7e27cf8 f7fd2ae0 c026cc58 f7759000 f7e27cf8 f7fd2ae0
010d7ce6
Call Trace:
[<c026fbc9>] deadline_merge+0x49/0x120
[<c026a438>] elv_merge+0x18/0x30
[<c026cc58>] __make_request+0xb8/0x440
[<c014dd96>] bio_alloc+0x126/0x1b0
[<c026d1ad>] generic_make_request+0x1cd/0x1e0
[<c026d239>] submit_bio+0x79/0x90
[<c014bd34>] __block_write_full_page+0x284/0x400
[<c014d342>] block_write_full_page+0x32/0xc0
[<c014f990>] blkdev_get_block+0x0/0x60
[<c017d55c>] ext3_writepage+0x25c/0x300
[<c014fae4>] blkdev_writepage+0x14/0x20
[<c014f990>] blkdev_get_block+0x0/0x60
[<c0166dd0>] mpage_writepages+0x1c0/0x2ab
[<c014fad0>] blkdev_writepage+0x0/0x20
[<c011ec8b>] do_softirq+0x5b/0xc0
[<c0150c11>] generic_writepages+0x11/0x15
[<c0134e46>] do_writepages+0x16/0x30
[<c0165548>] __sync_single_inode+0xc8/0x210
[<c01658aa>] sync_sb_inodes+0x16a/0x210
[<c01659a4>] writeback_inodes+0x54/0x90
[<c0134d05>] wb_kupdate+0xa5/0x100
[<c013537f>] __pdflush+0xff/0x1a0
[<c0135420>] pdflush+0x0/0x10
[<c013542b>] pdflush+0xb/0x10
[<c0134c60>] wb_kupdate+0x0/0x100
[<c0106fe0>] kernel_thread_helper+0x0/0x10
[<c0106fe5>] kernel_thread_helper+0x5/0x10
Code: 0f 0b 91 00 a0 f8 3c c0 8b 43 08 83 e0 7c 83 f8 10 75 09 8b
Janet Morgan <[email protected]> wrote:
>
> I got this on 2.5.63 while running the "find" command:
>
> kernel BUG at drivers/block/deadline-iosched.c:145!
> invalid operand: 0000
> CPU: 3
> EIP: 0060:[<c026f8a1>] Not tainted
> EFLAGS: 00010046
> EIP is at deadline_find_drq_hash+0x41/0xb0
> eax: f76fe060 ebx: f76fe058 ecx: f76fe060 edx: 00000000
> esi: f76fe060 edi: f76fe044 ebp: f76fe050 esp: f7e27c88
> ds: 007b es: 007b ss: 0068
> Process pdflush (pid: 16, threadinfo=f7e26000 task=c3b39300)
> Stack: 010d7ce6 00000000 f7fd2ae0 f7fd2ae0 f7fd2ae0 f7747f80 c026fbc9
> f7747f80
> 010d7ce6 00000000 f7747f80 00000000 f7759000 f7759000 00000008
> c026a438
> f7759000 f7e27cf8 f7fd2ae0 c026cc58 f7759000 f7e27cf8 f7fd2ae0
> 010d7ce6
> Call Trace:
> [<c026fbc9>] deadline_merge+0x49/0x120
> [<c026a438>] elv_merge+0x18/0x30
> [<c026cc58>] __make_request+0xb8/0x440
Nick, could this be caused by a stale next_drq[] entry?
If so,
http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.63/2.5.63-mm1/broken-out/deadline-dispatching-fix.patch
is designed to fix it.
Andrew Morton wrote:
> Janet Morgan <[email protected]> wrote:
> >
> > I got this on 2.5.63 while running the "find" command:
> >
> > kernel BUG at drivers/block/deadline-iosched.c:145!
> > invalid operand: 0000
> > CPU: 3
> > EIP: 0060:[<c026f8a1>] Not tainted
> > EFLAGS: 00010046
> > EIP is at deadline_find_drq_hash+0x41/0xb0
> > eax: f76fe060 ebx: f76fe058 ecx: f76fe060 edx: 00000000
> > esi: f76fe060 edi: f76fe044 ebp: f76fe050 esp: f7e27c88
> > ds: 007b es: 007b ss: 0068
> > Process pdflush (pid: 16, threadinfo=f7e26000 task=c3b39300)
> > Stack: 010d7ce6 00000000 f7fd2ae0 f7fd2ae0 f7fd2ae0 f7747f80 c026fbc9
> > f7747f80
> > 010d7ce6 00000000 f7747f80 00000000 f7759000 f7759000 00000008
> > c026a438
> > f7759000 f7e27cf8 f7fd2ae0 c026cc58 f7759000 f7e27cf8 f7fd2ae0
> > 010d7ce6
> > Call Trace:
> > [<c026fbc9>] deadline_merge+0x49/0x120
> > [<c026a438>] elv_merge+0x18/0x30
> > [<c026cc58>] __make_request+0xb8/0x440
>
> Nick, could this be caused by a stale next_drq[] entry?
>
> If so,
>
> http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.63/2.5.63-mm1/broken-out/deadline-dispatching-fix.patch
>
> is designed to fix it.
Thanks Andrew. 2.5.63-mm1 fixes the problem.
Andrew Morton wrote:
>Janet Morgan <[email protected]> wrote:
>
>
>>I got this on 2.5.63 while running the "find" command:
>>
>>kernel BUG at drivers/block/deadline-iosched.c:145!
>>invalid operand: 0000
>>CPU: 3
>>EIP: 0060:[<c026f8a1>] Not tainted
>>EFLAGS: 00010046
>>EIP is at deadline_find_drq_hash+0x41/0xb0
>>eax: f76fe060 ebx: f76fe058 ecx: f76fe060 edx: 00000000
>>esi: f76fe060 edi: f76fe044 ebp: f76fe050 esp: f7e27c88
>>ds: 007b es: 007b ss: 0068
>>Process pdflush (pid: 16, threadinfo=f7e26000 task=c3b39300)
>>Stack: 010d7ce6 00000000 f7fd2ae0 f7fd2ae0 f7fd2ae0 f7747f80 c026fbc9
>>f7747f80
>> 010d7ce6 00000000 f7747f80 00000000 f7759000 f7759000 00000008
>>c026a438
>> f7759000 f7e27cf8 f7fd2ae0 c026cc58 f7759000 f7e27cf8 f7fd2ae0
>>010d7ce6
>>Call Trace:
>> [<c026fbc9>] deadline_merge+0x49/0x120
>> [<c026a438>] elv_merge+0x18/0x30
>> [<c026cc58>] __make_request+0xb8/0x440
>>
>>
>
>Nick, could this be caused by a stale next_drq[] entry?
>
Yes, it could be.