Bug 706574

Summary: module wmi when unloaded causes Oops and system locks up
Product: [Fedora] Fedora Reporter: cam <camilo>
Component: kernelAssignee: John Feeney <jfeeney>
Status: CLOSED ERRATA QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 15CC: gansalmon, itamar, jonathan, kernel-maint, lekensteyn, madhu.chinakonda, sassmann
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2012-02-06 16:53:54 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
log messages from boot to eventual Oops
none
screenshot.jpg
none
Part of kernel log after Oops none

Description cam 2011-05-21 07:58:48 UTC 
Description of problem:
Whilst investigating problems related to https://bugzilla.redhat.com/show_bug.cgi?id=697157
I unloaded the hp_wmi module, then the wmi module. The system locked up, the screen changed to a plain console and an oops message was printed. I recovered this from the log after power cycling (the machine wasn't responding)

Version-Release number of selected component (if applicable):
kernel-2.6.38.6-27.fc15.i686


How reproducible:
100%

Steps to Reproduce:
1. boot to normal system
2. rmmod hp_wmi
3. rmmod wmi
  
Actual results:
black screen / console, oops message and system is locked up

Expected results:
module should be removable or give error message.

Additional info:
May be related to BZ 697157 where the ath9k module can't unload (and locks up spontaneously too)
Comment 1 cam 2011-05-21 08:10:51 UTC 
Created attachment 500172 [details]
log messages from boot to eventual Oops
Comment 2 Josh Boyer 2011-10-06 15:40:42 UTC 
Do you still get this error with the 2.6.40.x updates in F15?
Comment 3 Stefan Assmann 2011-10-07 10:38:50 UTC 
modprobe -r wmi on 2.6.40.6 still oopses.
Comment 4 Stefan Assmann 2011-10-17 07:40:16 UTC 
also still happens with f16 beta
Comment 5 Chuck Ebbert 2011-10-17 14:04:51 UTC 
(In reply to comment #4)
> also still happens with f16 beta

Please post traces from f15 and/or f16
Comment 6 Stefan Assmann 2011-10-17 14:49:26 UTC 
Created attachment 528559 [details]
screenshot.jpg

The system is inoperable after module unload so I could only take a screenshot.
f16 kernel-3.1.0.0.rc9
Comment 7 Peter Wu 2011-11-02 10:53:17 UTC 
Created attachment 531313 [details]
Part of kernel log after Oops

It happens on Kubuntu 11.10 Oneiric too

$ uname -a
Linux ubuntu 3.0.0-12-generic #20-Ubuntu SMP Fri Oct 7 14:56:25 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux

For some reason, device->kobj is NULL
sysfs_find_dirent: http://lxr.linux.no/#linux+v3.1/fs/sysfs/dir.c#L539
sysfs_get_dirent: http://lxr.linux.no/#linux+v3.1/fs/sysfs/dir.c#L575
sysfs_unmerge_group: http://lxr.linux.no/#linux+v3.1/fs/sysfs/group.c#L195
(rpm_sysfs_remove: http://lxr.linux.no/#linux+v3.1/drivers/base/power/sysfs.c#L534)
dpm_sysfs_remove: http://lxr.linux.no/#linux+v3.1/drivers/base/power/sysfs.c#L539
device_del: http://lxr.linux.no/#linux+v3.1/drivers/base/core.c#L1098
device_unregister: http://lxr.linux.no/#linux+v3.1/drivers/base/core.c#L1152

Could the patch at have fixed this?
http://git.kernel.org/?p=linux/kernel/git/next/linux-next-history.git;a=commitdiff;h=023b9565972a4a5e0f01b9aa32680af6e9b5c388
Comment 8 Josh Boyer 2011-11-02 13:04:40 UTC 
(In reply to comment #7)
> Could the patch at have fixed this?
> http://git.kernel.org/?p=linux/kernel/git/next/linux-next-history.git;a=commitdiff;h=023b9565972a4a5e0f01b9aa32680af6e9b5c388

Possibly.  That commit should be in the most recent rawhide kernel.  If someone is willing to test that, it would be appreciated.
Comment 9 Peter Wu 2011-11-02 13:44:20 UTC 
I've built the module using a custom Makefile and tested it. I can confirm that the NULL BUG() has disappeared.

Tests:

ubuntu@ubuntu:~$ sudo modprobe -v wmi 
insmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/wmi.ko 
ubuntu@ubuntu:~$ sudo modprobe -v -r wmi 
rmmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/wmi.ko
ubuntu@ubuntu:~$ sudo modprobe -v mxm-wmi 
insmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/wmi.ko 
insmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/mxm-wmi.ko 
ubuntu@ubuntu:~$ sudo modprobe -v -r mxm-wmi 
rmmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/mxm-wmi.ko
rmmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/wmi.ko
ubuntu@ubuntu:~$ sudo modprobe -v nouveau 
insmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/wmi.ko 
insmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/mxm-wmi.ko 
insmod /lib/modules/3.0.0-12-generic/kernel/drivers/gpu/drm/ttm/ttm.ko 
insmod /lib/modules/3.0.0-12-generic/kernel/drivers/gpu/drm/nouveau/nouveau.ko 
ubuntu@ubuntu:~$ sudo modprobe -v -r nouveau 
rmmod /lib/modules/3.0.0-12-generic/kernel/drivers/gpu/drm/nouveau/nouveau.ko
rmmod /lib/modules/3.0.0-12-generic/kernel/drivers/gpu/drm/ttm/ttm.ko
rmmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/mxm-wmi.ko
rmmod /lib/modules/3.0.0-12-generic/kernel/drivers/platform/x86/wmi.ko

No Oopses anymore.
Comment 10 Josh Boyer 2011-11-02 13:47:25 UTC 
(In reply to comment #9)
> I've built the module using a custom Makefile and tested it. I can confirm that
> the NULL BUG() has disappeared.

OK.  I'll backport the patch to F16 and get it included.  Will ship as an update, as it's a bit too late to get this into the release at this point.
Comment 11 Josh Boyer 2011-11-02 14:00:28 UTC 
Added to both F15 and F16.  Will be in the next kernel builds.

Also, this was supposed to head for stable so I pinged GregKH about it as well.