Red Hat Bugzilla – Bug 127862
INSTALL: kernel BUG at kernel/workqueue.c:104
Last modified: 2015-01-04 17:07:53 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7)
Description of problem:
Trying to install Fedora Core 3 test 1 on my Abit VP6-based system.
The following kernel BUG occurs right after I confirm my package
------------[ cut here ]------------
kernel BUG at kernel/workqueue.c:104!
invalid operand: 0000 [#2]
Modules linked in: parport_pc parport dm_mod xfs jfs reiserfs ext3 jbd
d6 raid5 xor raid1 raid0 8139too mii uhci_hcd sr_mod sd_mod scsi_mod
nfs lockd sunrpc vfat fat cramfs vga16fb vgastate
EIP: 0060:[<c0130cb4>] Not tainted
EFLAGS: 00210202 (2.6.7-1.478)
EIP is at queue_work+0x17/0x2f
eax: c2187620 ebx: f7d22148 ecx: 00000000 edx: c218761c
esi: c02328e6 edi: c03424a0 ebp: c03bffd0 esp: c03bffc8
ds: 007b es: 007b ss: 0068
Process anaconda (pid: 429, threadinfo=c03bf000 task=f6fef970)
Stack: c21874e4 c012831f c03bffd0 c03bffd0 00000000 00000011 c03e4ca8
00000000 c0124185 f6c15e44 00200046 c03c0000 c0108d17
Stack pointer is garbage, not printing trace
Code: 0f 0b 68 00 3f 8a 30 c0 89 d8 e8 e5 fe ff ff b9 01 00 00 00
<0>Kernel panic: Fatal exception in interrupt
In interrupt handler - not syncing
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Attempt to install Fedora Core 3 test 1 on an Abit VP6-based system.
Created attachment 104917 [details]
FC3t2 installation BUG test
This bug is still present in test 2.
I've managed to work around this BUG by disabling the drives that
contain my software RAID partitions (hde=none hdg=none hdi=none
hdk=none) and installing a minimal system on hda. hda is attached to
the built in VIA controller on the VP6; hde-hdk are attached to a
couple of Maxtor-branded Promise PDC20269 (Ultra133TX2) controllers.
Note that simply installing to a partition on /dev/hda wasn't
sufficient, possibly because of the swap "filesystem" on /dev/md2.
It's only a guess, but I think that the BUG may be triggered when
Anaconda tries to start a software RAID device. queue_work seems to
be involved in starting new kernel threads, and this is the most only
reason I can come up with that a new kernel thread would be created at
this point in the installation process.
Still there in test 3 also.
Still present in RC1.
Still present in RC2.
is it the exact same trace in RC2 ? No further information, like a stack trace ?
Created attachment 105866 [details]
BUG output from FC3rc2
Created attachment 105869 [details]
serial console capture from FC3rc2
I got lazy and used a serial console to capture the BUG from FC3rc2. Scanning
through the output I noticed that there is a kernel OOPS which I never saw
before. It occurs during the "Searching for Fedora Core installations..."
Please try kernel-2.6.9-1.649 - it supposedly has a fix for the oops
*** Bug 137273 has been marked as a duplicate of this bug. ***
Where can I find 2.6.9-1.649?
Never mind. RC5 successfully installs -- no OOPS or BUG.