Re: 2.6.0-test4: blkdev_requests "memory before object was overwritten" and oops in __iget

From: Oleg Drokin
Date: Thu Aug 28 2003 - 05:35:56 EST

Next message: Mirko Lindner: "[PATCH 2.6.0-test4][sk98lin] sk98lin driver with hardware bug makeeth unusable"
Previous message: Timo Sirainen: "RE: Lockless file reading"
In reply to: Andrew Morton: "Re: 2.6.0-test4: blkdev_requests "memory before object wasoverwritten" and oops in __iget"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hello!

On Tue, Aug 26, 2003 at 02:36:35PM -0700, Andrew Morton wrote:
> Oleg Drokin <green@xxxxxxxxxxxxxx> wrote:
> > Found one of my test boxes freezed today (or may be even yesterday).
> > That's what was in logs:
> >
> > Aug 25 00:50:37 dwarf kernel: [drm] DMA Cleanup
> > Aug 25 21:46:31 dwarf kernel: mtrr: base(0xf8000000) is not aligned on a size(0x180000) boundary
> > Aug 25 21:46:31 dwarf kernel: [drm] Using POST v1.2 init.
> > Aug 25 21:46:31 dwarf kernel: PCI: Found IRQ 11 for device 0000:00:02.0
> > Aug 25 21:46:39 dwarf kernel: slab error in cache_alloc_debugcheck_after(): cache `blkdev_requests': memory before object was overwritten
> Setting CONFIG_DEBUG_PAGEALLOC might help us trap this.

Actually I usually compile with this option, but not that time.
Well, so I run with CONFIG_DEBUG_PAGEALLOC and after two more nights (seems it
only crashes after two nights sitting idle) I got this:

PCI: Found IRQ 11 for device 0000:00:02.0
[drm] DMA Cleanup

I tried to start X at this point

mtrr: base(0xf8000000) is not aligned on a size(0x180000) boundary
[drm] Using POST v1.2 init.
PCI: Found IRQ 11 for device 0000:00:02.0
[drm:i810_wait_ring] *ERROR* space: 65520 wanted 65528
[drm:i810_wait_ring] *ERROR* lockup

And now X server segfaulted at this point, but no oops yet.

So I tried to start it again

mtrr: base(0xf8000000) is not aligned on a size(0x180000) boundary
[drm] Using POST v1.2 init.
PCI: Found IRQ 11 for device 0000:00:02.0
Unable to handle kernel NULL pointer dereference at virtual address 00000000
printing eip:
c018ecfd
*pde = 00000000
Oops: 0002 [#1]
CPU: 0
EIP: 0060:[<c018ecfd>] Not tainted
EFLAGS: 00010206
EIP is at prune_dcache+0x9d/0x640
eax: c033f544 ebx: cc3d7028 ecx: 000000d0 edx: 00000000
esi: d3184004 edi: 00000015 ebp: df975e80 esp: df975e58
ds: 007b es: 007b ss: 0068
Process kswapd0 (pid: 8, threadinfo=df974000 task=df9a5000)
Stack: 0000010b 0000010b df975e60 df975e60 00000000 00000029 df975e70 00000080
df974000 00000270 df975e90 c018ff0e 00000080 00000080 df975ec4 c01579ff
00000080 000000d0 00007819 01d90d30 00000000 000003f0 00000000 dfebed08
Call Trace:
[<c018ff0e>] shrink_dcache_memory+0x3e/0x40
[<c01579ff>] shrink_slab+0x11f/0x170
[<c0159850>] balance_pgdat+0x1e0/0x220
[<c0159975>] kswapd+0xe5/0x100
[<c0123ac0>] autoremove_wake_function+0x0/0x50
[<c0123ac0>] autoremove_wake_function+0x0/0x50
[<c0159890>] kswapd+0x0/0x100
[<c0108269>] kernel_thread_helper+0x5/0xc

Code: 89 02 a1 48 f5 33 c0 89 1b 0f 18 00 90 ff 0d 50 f5 33 c0 8d
lib/dec_and_lock.c:32: spin_lock(fs/dcache.c:c033f510) already locked by fs/dcache.c/369

And other oopses that are probably because of first one.
Looks like some memory corruption to me, but the memory on the box is fine.

Unable to handle kernel NULL pointer dereference at virtual address 00000000
printing eip:
c018fb9f
*pde = 00000000
Oops: 0000 [#2]
CPU: 0
EIP: 0060:[<c018fb9f>] Not tainted
EFLAGS: 00010246
EIP is at select_parent+0xcf/0x200
eax: 00000000 ebx: d04cd004 ecx: d04cd028 edx: d9bb2028
esi: dde05038 edi: dde05038 ebp: d0b2bec0 esp: d0b2be90
ds: 007b es: 007b ss: 0068
Process kdeinit (pid: 6132, threadinfo=d0b2a000 task=d57bd000)
Stack: 00000001 00000000 d0b2bed0 00000086 c10b8dd0 00000001 00000000 00000000
dde05004 c770e004 c770e004 000017f5 d0b2bed0 c018fce8 c770e004 c770e004
d0b2bee0 c01ad658 c770e004 dd3ab000 d0b2bf0c c0128108 c770e004 dfebd940
Call Trace:
[<c018fce8>] shrink_dcache_parent+0x18/0x30
[<c01ad658>] proc_pid_flush+0x18/0x30
[<c0128108>] release_task+0x1c8/0x2d0
[<c0120b57>] schedule+0x2c7/0x530
[<c012a710>] wait_task_zombie+0x130/0x190
[<c012aaea>] sys_wait4+0x20a/0x250
[<c0120dc0>] default_wake_function+0x0/0x30
[<c0120dc0>] default_wake_function+0x0/0x30
[<c012ab57>] sys_waitpid+0x27/0x2b
[<c010aedd>] sysenter_past_esp+0x52/0x71

Code: 8b 10 89 4a 04 89 53 24 89 41 04 89 08 ff 05 50 f5 33 c0 ff
lib/dec_and_lock.c:32: spin_lock(fs/dcache.c:c033f510) already locked by fs/dcache.c/533
Unable to handle kernel NULL pointer dereference at virtual address 00000000
printing eip:
c018fb9f
*pde = 00000000
Oops: 0000 [#3]
CPU: 0
EIP: 0060:[<c018fb9f>] Not tainted
EFLAGS: 00010246
EIP is at select_parent+0xcf/0x200
eax: 00000000 ebx: c23ee004 ecx: c23ee028 edx: c033f544
esi: d2b36038 edi: d2b36038 ebp: d0b2bec0 esp: d0b2be90
ds: 007b es: 007b ss: 0068
Process dcopserver_shut (pid: 6134, threadinfo=d0b2a000 task=d57bd000)
Stack: 00000001 00000000 d0b2bed0 00000086 c10b8dd0 00000001 00000000 00000000
d2b36004 cecd9004 cecd9004 000017f7 d0b2bed0 c018fce8 cecd9004 cecd9004
d0b2bee0 c01ad658 cecd9004 c80f7000 d0b2bf0c c0128108 cecd9004 cb609004
Call Trace:
[<c018fce8>] shrink_dcache_parent+0x18/0x30
[<c01ad658>] proc_pid_flush+0x18/0x30
[<c0128108>] release_task+0x1c8/0x2d0
[<c0120a8a>] schedule+0x1fa/0x530
[<c012a710>] wait_task_zombie+0x130/0x190
[<c012aaea>] sys_wait4+0x20a/0x250
[<c0120dc0>] default_wake_function+0x0/0x30
[<c0120dc0>] default_wake_function+0x0/0x30
[<c012ab57>] sys_waitpid+0x27/0x2b
[<c010aedd>] sysenter_past_esp+0x52/0x71

Code: 8b 10 89 4a 04 89 53 24 89 41 04 89 08 ff 05 50 f5 33 c0 ff
lib/dec_and_lock.c:32: spin_lock(fs/dcache.c:c033f510) already locked by fs/dcache.c/533
Unable to handle kernel NULL pointer dereference at virtual address 00000000
printing eip:
c018fb9f
*pde = 00000000
Oops: 0000 [#4]
CPU: 0
EIP: 0060:[<c018fb9f>] Not tainted
EFLAGS: 00010246
EIP is at select_parent+0xcf/0x200
eax: 00000000 ebx: d526e004 ecx: d526e028 edx: cc94f028
esi: db032038 edi: db032038 ebp: cbaa9ec0 esp: cbaa9e90
ds: 007b es: 007b ss: 0068
Process kdeinit (pid: 6130, threadinfo=cbaa8000 task=c336c000)
Stack: 00000001 00000000 cbaa9ed0 00000086 c14d1ea8 00000001 00000000 00000000
db032004 c6b5b004 c6b5b004 000017f6 cbaa9ed0 c018fce8 c6b5b004 c6b5b004
cbaa9ee0 c01ad658 c6b5b004 d57bd000 cbaa9f0c c0128108 c6b5b004 cb609004
Call Trace:
[<c018fce8>] shrink_dcache_parent+0x18/0x30
[<c01ad658>] proc_pid_flush+0x18/0x30
[<c0128108>] release_task+0x1c8/0x2d0
[<c0120a8a>] schedule+0x1fa/0x530
[<c012a710>] wait_task_zombie+0x130/0x190
[<c012aaea>] sys_wait4+0x20a/0x250
[<c0120dc0>] default_wake_function+0x0/0x30
[<c0120dc0>] default_wake_function+0x0/0x30
[<c012ab57>] sys_waitpid+0x27/0x2b
[<c010aedd>] sysenter_past_esp+0x52/0x71

Code: 8b 10 89 4a 04 89 53 24 89 41 04 89 08 ff 05 50 f5 33 c0 ff
lib/dec_and_lock.c:32: spin_lock(fs/dcache.c:c033f510) already locked by fs/dcache.c/533
Unable to handle kernel NULL pointer dereference at virtual address 00000000
printing eip:
c018fb9f
*pde = 00000000
Oops: 0000 [#5]
CPU: 0
EIP: 0060:[<c018fb9f>] Not tainted
EFLAGS: 00010246
EIP is at select_parent+0xcf/0x200
eax: 00000000 ebx: cfc98004 ecx: cfc98028 edx: c033f544
esi: da1fe038 edi: da1fe038 ebp: cfa97ec0 esp: cfa97e90
ds: 007b es: 007b ss: 0068
Process kdeinit (pid: 6142, threadinfo=cfa96000 task=cc3e4000)
Stack: 00000001 00000000 cfa97ed0 00000086 c13837b8 00000001 00000000 00000000
da1fe004 cf59e004 cf59e004 000017ff cfa97ed0 c018fce8 cf59e004 cf59e004
cfa97ee0 c01ad658 cf59e004 df9a5000 cfa97f0c c0128108 cf59e004 d6d0fbfc
Call Trace:
[<c018fce8>] shrink_dcache_parent+0x18/0x30
[<c01ad658>] proc_pid_flush+0x18/0x30
[<c0128108>] release_task+0x1c8/0x2d0
[<c012a710>] wait_task_zombie+0x130/0x190
[<c012aaea>] sys_wait4+0x20a/0x250
[<c0120dc0>] default_wake_function+0x0/0x30
[<c01261a6>] __mmdrop+0x36/0x47
[<c0120dc0>] default_wake_function+0x0/0x30
[<c012ab57>] sys_waitpid+0x27/0x2b
[<c010aedd>] sysenter_past_esp+0x52/0x71

Code: 8b 10 89 4a 04 89 53 24 89 41 04 89 08 ff 05 50 f5 33 c0 ff
lib/dec_and_lock.c:32: spin_lock(fs/dcache.c:c033f510) already locked by fs/dcache.c/533
Unable to handle kernel NULL pointer dereference at virtual address 00000000
printing eip:
c018fb9f
*pde = 00000000
Oops: 0000 [#6]
CPU: 0
EIP: 0060:[<c018fb9f>] Not tainted
EFLAGS: 00010246
EIP is at select_parent+0xcf/0x200
eax: 00000000 ebx: c5b9e004 ecx: c5b9e028 edx: d68f7028
esi: c89f9038 edi: c89f9038 ebp: cfa97ec0 esp: cfa97e90
ds: 007b es: 007b ss: 0068
Process dcopserver_shut (pid: 6144, threadinfo=cfa96000 task=cc3e4000)
Stack: 00000001 00000000 cfa97ed0 00000086 c13837b8 00000001 00000000 00000000
c89f9004 c4124004 c4124004 00001801 cfa97ed0 c018fce8 c4124004 c4124004
cfa97ee0 c01ad658 c4124004 c69b3000 cfa97f0c c0128108 c4124004 d05cdbfc
Call Trace:
[<c018fce8>] shrink_dcache_parent+0x18/0x30
[<c01ad658>] proc_pid_flush+0x18/0x30
[<c0128108>] release_task+0x1c8/0x2d0
[<c012a710>] wait_task_zombie+0x130/0x190
[<c012aaea>] sys_wait4+0x20a/0x250
[<c0120dc0>] default_wake_function+0x0/0x30
[<c01261a6>] __mmdrop+0x36/0x47
[<c0120dc0>] default_wake_function+0x0/0x30
[<c012ab57>] sys_waitpid+0x27/0x2b
[<c010aedd>] sysenter_past_esp+0x52/0x71

Code: 8b 10 89 4a 04 89 53 24 89 41 04 89 08 ff 05 50 f5 33 c0 ff
lib/dec_and_lock.c:32: spin_lock(fs/dcache.c:c033f510) already locked by fs/dcache.c/533
Unable to handle kernel NULL pointer dereference at virtual address 00000000
printing eip:
c018fb9f
*pde = 00000000
Oops: 0000 [#7]
CPU: 0
EIP: 0060:[<c018fb9f>] Not tainted
EFLAGS: 00010246
EIP is at select_parent+0xcf/0x200
eax: 00000000 ebx: c4b0c004 ecx: c4b0c028 edx: d68f7028
esi: da14e038 edi: da14e038 ebp: deda5ec0 esp: deda5e90
ds: 007b es: 007b ss: 0068
Process kdeinit (pid: 6140, threadinfo=deda4000 task=da608000)
Stack: 00000001 00000000 deda5ed0 00000086 c13967f0 00000001 00000000 00000000
da14e004 d3919004 d3919004 00001800 deda5ed0 c018fce8 d3919004 d3919004
deda5ee0 c01ad658 d3919004 cc3e4000 deda5f0c c0128108 d3919004 cb609004
Call Trace:
[<c018fce8>] shrink_dcache_parent+0x18/0x30
[<c01ad658>] proc_pid_flush+0x18/0x30
[<c0128108>] release_task+0x1c8/0x2d0
[<c0120a8a>] schedule+0x1fa/0x530
[<c012a710>] wait_task_zombie+0x130/0x190
[<c012aaea>] sys_wait4+0x20a/0x250
[<c0120dc0>] default_wake_function+0x0/0x30
[<c0120dc0>] default_wake_function+0x0/0x30
[<c012ab57>] sys_waitpid+0x27/0x2b
[<c010aedd>] sysenter_past_esp+0x52/0x71

Code: 8b 10 89 4a 04 89 53 24 89 41 04 89 08 ff 05 50 f5 33 c0 ff
lib/dec_and_lock.c:32: spin_lock(fs/dcache.c:c033f510) already locked by fs/dcache.c/533

Bye,
Oleg
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Mirko Lindner: "[PATCH 2.6.0-test4][sk98lin] sk98lin driver with hardware bug makeeth unusable"
Previous message: Timo Sirainen: "RE: Lockless file reading"
In reply to: Andrew Morton: "Re: 2.6.0-test4: blkdev_requests "memory before object wasoverwritten" and oops in __iget"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]