kernel 2.6.20-15-generic oops in ext3
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
linux (Ubuntu) |
Invalid
|
Undecided
|
Unassigned | ||
linux-source-2.6.20 (Ubuntu) |
Won't Fix
|
Medium
|
Unassigned |
Bug Description
Binary package hint: linux-image-
I've had the kernel oops and hang the machine in the last kernels.
I've had the machine's ram, cpu and disks tested for a week with no errors found.
I had to reinstall Ubuntu, so I used the new 7.04 for i386, with kernel 2.6.20-15-generic
The machine is an AMD Athlon(tm) 64 Processor 3200+
This oops was generated after some hours of running.
I had to copy it by hand from the monitor, since the machine was hung and no copy was in /var/log/messages.
I don't have any other machines with serial ports, so I cannot get the full oops.
I installed the kernel-debug package, but I couldn't find any documentation on how to use it to narrow the error down.
I couldn't find any similar looking oops either in the ubuntu bug list or lkml (I googled the names of the functions in the backtrace, but nothing caught my eye)
I'm running some heavy reading and writing on the disks, so I suspect locking errors leading to corruption (wild guess)
If you need more info (lspci, modules, etc), I'll post it.
The oops follows:
-------
ext3_journal_
ext3_ordered_
generic_
ext3_mark_
current_fs_time +0x50/0x60
__generic_
find_extend_vma +0x1d/0x70
get_futex_key +0x40/0x110
try_to_
generic_
ext3_file_write +0x30/0xc0 [ext3]
do_sync_write +0xd5/0x120
autoremove_
vfs_write +0xbe/0x190
sys_futex +0x91/0x140
do_sync_write +0x0/0x120
sys_write +0x41/0x70
sysenter_
Code:
F4 89 55 F0 89 75 F8 BE 40 6E 42 10 9C 58 FA 66 66 66 90 66 66 66 90 66 66 66 90 66 66 90 8B 55 F0
89 F3 89 02 8B 47 04 8B 40 10 <03> 1C 85 00 FC 3C C0 89 D8 E8 DA FF 1C 00 8B 47 04 8B 40 10 8B
EIP:[<C011E718>] task_rq_lock + 0x38 / 0x80
SS:ESP 0068:F029FBB0
<0> Kernel panic - not syncing: Fatal exception in interrupt.
------
LEDS:
Num_lock OFF
Caps_lock BLINKING
Scroll_lock BLINKING
I can reproduce the oops by letting my machine run for some days; the oops eventually hangs the machine.
However, I do not know a way to trigger it easily.