Out of nowhere on mozilla open (after it worked fine all afternoon):
------------[ cut here ]------------
kernel BUG at include/linux/blkdev.h:408!
invalid operand: 0000 [#1]
CPU: 0
EIP: 0060:[<c02322be>] Tainted: P
EFLAGS: 00010046
EIP is at blk_queue_start_tag+0x8e/0x100
eax: dfdfaf40 ebx: 00000000 ecx: c040176c edx: dfdfc1a8
esi: dfdfc1a8 edi: dff52240 ebp: dff52260 esp: cf275aa8
ds: 007b es: 007b ss: 0068
Process mozilla-bin (pid: 3073, threadinfo=cf274000 task=d7296cc0)
Stack: da5a3880 000003e8 c04016c0 c040176c 00000286 00000001 dfdf0080 c040176c
c0245baa c040177c dfdfc1a8 00e916ff c04016c0 c0246832 c040176c dfdfc1a8
c01142d4 00000060 00000283 00000000 c01e92f2 000003e8 0023e546 000001f7
Call Trace:
[<c0245baa>] idedisk_start_tag+0x7a/0x90
[<c0246832>] __ide_do_rw_disk+0x722/0x770
[<c01142d4>] delay_tsc+0x14/0x40
[<c01e92f2>] __delay+0x12/0x20
[<c023afb4>] start_request+0x104/0x160
[<c023b28b>] ide_do_request+0x24b/0x430
[<c023b48d>] do_ide_request+0x1d/0x30
[<c0233c5b>] __make_request+0x2bb/0x4f0
[<c0233fb1>] generic_make_request+0x121/0x1b0
[<c011e1d0>] autoremove_wake_function+0x0/0x50
[<c023407d>] submit_bio+0x3d/0x70
[<c0173a5e>] mpage_bio_submit+0x2e/0x40
[<c0173d22>] do_mpage_readpage+0x182/0x320
[<c0155347>] __find_get_block+0x67/0xe0
[<c01391e0>] add_to_page_cache+0x70/0x110
[<c0173feb>] mpage_readpages+0x12b/0x160
[<c01bd790>] reiserfs_get_block+0x0/0x15a0
[<c013cb3f>] __rmqueue+0xbf/0x110
[<c013eca5>] read_pages+0x115/0x120
[<c01bd790>] reiserfs_get_block+0x0/0x15a0
[<c013d072>] __alloc_pages+0x92/0x320
[<c013eded>] do_page_cache_readahead+0x13d/0x1c0
[<c013efb5>] page_cache_readahead+0x145/0x190
[<c0139a42>] do_generic_mapping_read+0xb2/0x3b0
[<c0139d40>] file_read_actor+0x0/0x110
[<c013a009>] __generic_file_aio_read+0x1b9/0x200
[<c0139d40>] file_read_actor+0x0/0x110
[<c013a15e>] generic_file_read+0x8e/0xb0
[<c01609a0>] link_path_walk+0x610/0x8e0
[<c015fd50>] vfs_permission+0x80/0x110
[<c015fe34>] permission+0x54/0x70
[<c016139f>] may_open+0x5f/0x230
[<c0153600>] get_empty_filp+0x90/0x100
[<c016160d>] open_namei+0x9d/0x440
[<c0151a0a>] dentry_open+0xea/0x210
[<c01c18c0>] reiserfs_file_release+0x0/0x490
[<c0169962>] dput+0x22/0x1e0
[<c01c18c0>] reiserfs_file_release+0x0/0x490
[<c0152702>] vfs_read+0xe2/0x150
[<c01522e0>] default_llseek+0x0/0xd0
[<c01529c2>] sys_read+0x42/0x70
[<c01092a5>] sysenter_past_esp+0x52/0x71
Code: 0f 0b 98 01 6e 51 32 c0 eb b6 8b 4f 14 90 8d 74 26 00 83 c3
<6>note: mozilla-bin[3073] exited with preempt_count 1
On Tue, May 27, 2003 at 11:23:29PM -0400, Ivan Gyurdiev wrote:
> Out of nowhere on mozilla open (after it worked fine all afternoon):
> ------------[ cut here ]------------
> kernel BUG at include/linux/blkdev.h:408!
> invalid operand: 0000 [#1]
> CPU: 0
> EIP: 0060:[<c02322be>] Tainted: P
> EFLAGS: 00010046
> EIP is at blk_queue_start_tag+0x8e/0x100
This appears to be tainted by a proprietary module. Please reproduce
without it or forward the bugreport to the originator of the module.
-- wli
On Tue, 27 May 2003, William Lee Irwin III wrote:
> On Tue, May 27, 2003 at 11:23:29PM -0400, Ivan Gyurdiev wrote:
> > Out of nowhere on mozilla open (after it worked fine all afternoon):
> > ------------[ cut here ]------------
> > kernel BUG at include/linux/blkdev.h:408!
> > invalid operand: 0000 [#1]
> > CPU: 0
> > EIP: 0060:[<c02322be>] Tainted: P
> > EFLAGS: 00010046
> > EIP is at blk_queue_start_tag+0x8e/0x100
>
> This appears to be tainted by a proprietary module. Please reproduce
> without it or forward the bugreport to the originator of the module.
Looks valid;
We tried to remove the previous request.. but there was none. The BUG
check looks odd (what happens to the first tag?)
Zwane
--
function.linuxpower.ca
On Tue, May 27, 2003 at 11:23:29PM -0400, Ivan Gyurdiev wrote:
>>> Out of nowhere on mozilla open (after it worked fine all afternoon):
>>> ------------[ cut here ]------------
>>> kernel BUG at include/linux/blkdev.h:408!
>>> invalid operand: 0000 [#1]
>>> CPU: 0
>>> EIP: 0060:[<c02322be>] Tainted: P
>>> EFLAGS: 00010046
>>> EIP is at blk_queue_start_tag+0x8e/0x100
On Tue, 27 May 2003, William Lee Irwin III wrote:
>> This appears to be tainted by a proprietary module. Please reproduce
>> without it or forward the bugreport to the originator of the module.
On Wed, May 28, 2003 at 01:38:35AM -0400, Zwane Mwaikambo wrote:
> Looks valid;
> We tried to remove the previous request.. but there was none. The BUG
> check looks odd (what happens to the first tag?)
Okay, I'll back off and let someone with a remote clue of what's going
on in ll_rw_blk.c take over.
-- wli
On Tue, May 27 2003, Ivan Gyurdiev wrote:
> Out of nowhere on mozilla open (after it worked fine all afternoon):
>
> ------------[ cut here ]------------
> kernel BUG at include/linux/blkdev.h:408!
> invalid operand: 0000 [#1]
> CPU: 0
> EIP: 0060:[<c02322be>] Tainted: P
> EFLAGS: 00010046
> EIP is at blk_queue_start_tag+0x8e/0x100
> eax: dfdfaf40 ebx: 00000000 ecx: c040176c edx: dfdfc1a8
> esi: dfdfc1a8 edi: dff52240 ebp: dff52260 esp: cf275aa8
> ds: 007b es: 007b ss: 0068
> Process mozilla-bin (pid: 3073, threadinfo=cf274000 task=d7296cc0)
> Stack: da5a3880 000003e8 c04016c0 c040176c 00000286 00000001 dfdf0080 c040176c
> c0245baa c040177c dfdfc1a8 00e916ff c04016c0 c0246832 c040176c dfdfc1a8
> c01142d4 00000060 00000283 00000000 c01e92f2 000003e8 0023e546 000001f7
> Call Trace:
> [<c0245baa>] idedisk_start_tag+0x7a/0x90
> [<c0246832>] __ide_do_rw_disk+0x722/0x770
> [<c01142d4>] delay_tsc+0x14/0x40
> [<c01e92f2>] __delay+0x12/0x20
> [<c023afb4>] start_request+0x104/0x160
> [<c023b28b>] ide_do_request+0x24b/0x430
> [<c023b48d>] do_ide_request+0x1d/0x30
This bug doesn't look possible. Basically it's a trigger to sanity check
whether the request is already gone from the list or not, and there's no
way it can be gone from this path (IDE would never call
idedisk_start_tag() on a removed request).
So I suspect something else is wrong, and this was just the first to
trigger.
--
Jens Axboe
On Wed, May 28 2003, Zwane Mwaikambo wrote:
> On Tue, 27 May 2003, William Lee Irwin III wrote:
>
> > On Tue, May 27, 2003 at 11:23:29PM -0400, Ivan Gyurdiev wrote:
> > > Out of nowhere on mozilla open (after it worked fine all afternoon):
> > > ------------[ cut here ]------------
> > > kernel BUG at include/linux/blkdev.h:408!
> > > invalid operand: 0000 [#1]
> > > CPU: 0
> > > EIP: 0060:[<c02322be>] Tainted: P
> > > EFLAGS: 00010046
> > > EIP is at blk_queue_start_tag+0x8e/0x100
> >
> > This appears to be tainted by a proprietary module. Please reproduce
> > without it or forward the bugreport to the originator of the module.
>
> Looks valid;
>
> We tried to remove the previous request.. but there was none. The BUG
> check looks odd (what happens to the first tag?)
Previous request, what are you talking about?
do_ide_request()
ide_do_request()
start_request(drive, rq); /* this rq is _never_ off the list */
__ide_do_rw_disk(drive, rq);
idedisk_start_tag(drive, rq);
blkdev_dequeue_request(rq); /* boom */
It does _not_ look valid.
--
Jens Axboe
On Wed, 28 May 2003, Jens Axboe wrote:
> Previous request, what are you talking about?
I meant the rq will be removed off the request queue.
blk_queue_start_tag
blkdev_dequeue_request;
Bad wording, clue anaemia.
> do_ide_request()
> ide_do_request()
> start_request(drive, rq); /* this rq is _never_ off the list */
> __ide_do_rw_disk(drive, rq);
> idedisk_start_tag(drive, rq);
> blkdev_dequeue_request(rq); /* boom */
>
> It does _not_ look valid.
And thus he spake...
--
function.linuxpower.ca
On Wed, May 28 2003, Zwane Mwaikambo wrote:
> On Wed, 28 May 2003, Jens Axboe wrote:
>
> > Previous request, what are you talking about?
>
> I meant the rq will be removed off the request queue.
>
> blk_queue_start_tag
> blkdev_dequeue_request;
Yes that is fine, the problem is that blkdev_dequeue_request() complains
about the request not being on the list. Which looks very corrupt.
--
Jens Axboe