Red Hat Bugzilla – Bug 190360
kernel panic on bad locking
Last modified: 2007-11-30 17:11:31 EST
Description of problem:
Booting the kernel in question causes a kernel panic with a reason of "bad locking".
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. boot kernel in question
Welcome to Fedora Core
Press 'I' to enter interactive startup.
Setting clock (localtime): Mon May 1 10:08:29 MDT 2006 [ OK ]
Starting udev: BUG: spinlock bad magic on CPU#0, events/0/5 (Not tainted)
lock: dfd16200, .magic: 00000000, .owner: <none>/-1, .owner_cpu: 0
Kernel panic - not syncing: bad locking
Even though the output seems to imply the bug is in udev, I'm filing against
kernel because booting an earlier kernel (kernel-2.6.16-1.2147_FC6) does not
show this problem.
I encountered the same problem.
In my case, the bad locking bug was caused by bcm43xx driver. When I add the
bcm43xx to /etc/modprobe.d/blacklist, the udev does not cause the panic.
I dug deep into this bug, and I found that it was caused by netif_tx_disable()
in the ieee80211softmac_start_scan(). In the netif_tx_disable(), dev->xmit_lock
was locked, but it was not initialized yet.
I think the linux-2.6-bcm43xx-assoc-on-startup.patch was caused this bug, because
this patch called schedule_work(&bcm->softmac->associnfo.work) before calling
register_netdev() function, but the dev->xmit_lock was initialized in the
Fixed in recent rawhide kernels.