Bug 144041

Summary: kernel oopses when disconnecting usb modem (cdc-acm.ko)
Product: [openSUSE] SUSE Linux 10.1 Reporter: Dirk Mueller <dmueller>
Component: KernelAssignee: Greg Kroah-Hartman <gregkh>
Status: VERIFIED FIXED QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P5 - None CC: vojtech
Version: Beta 1   
Target Milestone: ---   
Hardware: Other   
OS: Other   
Whiteboard:
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: oops
2nd oops
3rd oops
4th oops
5th oops
6th oops
/proc/bus/usb/devices
7th oops
Potential patch to fix the cdc-acm oops

Description Dirk Mueller 2006-01-19 10:38:57 UTC
kernel oopses when disconnecting usb modem
Comment 1 Dirk Mueller 2006-01-19 10:39:22 UTC
Created attachment 63948 [details]
oops
Comment 2 Forgotten User ZhJd0F0L3x 2006-01-19 13:24:47 UTC
might be related to bug #143408
Comment 3 Dirk Mueller 2006-01-19 13:30:19 UTC
hmm, could be, but the oops is different. lets see what the $guru says. 
Comment 4 Greg Kroah-Hartman 2006-01-19 18:15:09 UTC
Ugh, this should have been fixed already.

Does this happen in 2.6.16-rc1?
Comment 5 Dirk Mueller 2006-01-19 21:34:54 UTC
I don't know, I used the most recent version in autobuild. is there a newer kernel build available?

its 2.6.15-git12-5-smp where I can reproduce the bug. 

Comment 6 Dirk Mueller 2006-01-30 12:22:02 UTC
it still happens with 2.6.16-rc1-git3-4-default
Comment 7 Dirk Mueller 2006-01-30 12:22:33 UTC
Created attachment 65621 [details]
2nd oops
Comment 8 Dirk Mueller 2006-02-01 08:39:13 UTC
Created attachment 65981 [details]
3rd oops

still oopses
Comment 9 Dirk Mueller 2006-02-06 19:49:31 UTC
Created attachment 66594 [details]
4th oops

still oopses
Comment 10 Forgotten User ZhJd0F0L3x 2006-02-06 20:52:05 UTC
I'll add two of the original authors.
Comment 11 Greg Kroah-Hartman 2006-02-06 22:01:48 UTC
I have a potential patch that fixes this, are you able to apply it and
rebuild a kernel?

If so, it's at:
  http://www.kernel.org/pub/linux/kernel/people/gregkh/gregkh-2.6/patches/usb/usb-fix-oops-in-acm-disconnect.patch

if not, please let me know.
Comment 12 Dirk Mueller 2006-02-06 22:07:53 UTC
wow, this looks like the one I'm looking for. 

no, not easily. current kernel-default doesn't even build without the patch ;(

Comment 13 Greg Kroah-Hartman 2006-02-06 22:29:34 UTC
What kernel source version did you try to apply it against?
Comment 14 Dirk Mueller 2006-02-07 20:22:39 UTC
well, the one from stable. it doesn't apply because it is apparently applied already. 
Comment 15 Dirk Mueller 2006-02-07 20:23:24 UTC
Created attachment 66870 [details]
5th oops

still oopses
Comment 16 Dirk Mueller 2006-02-07 20:24:07 UTC
Created attachment 66871 [details]
6th oops

still oopses
Comment 17 Greg Kroah-Hartman 2006-02-07 20:35:43 UTC
No, that patch should not be already applied to any suse kernel.

What specific version number did you try to patch it with?
Comment 18 Dirk Mueller 2006-02-07 20:41:52 UTC
the one from stable. it identifies itself as 2.6.16-rc2-git2-3-default
Comment 19 Dirk Mueller 2006-02-07 20:43:11 UTC
see /work/built/mbuild/oldboy-dmueller/1514/i386. and indeed, looking at it with human eyes and with md5 sum the patch you posted is in patches.kernel.org/patch-2.6.16-rc1-rc2
Comment 20 Dirk Mueller 2006-02-07 20:44:27 UTC
logfile says: 

+ patch -s -E -p1 --no-backup-if-mismatch -i ../patches.fixes/usb-fix-oops-in-acm-disconnect.patch
Reversed (or previously applied) patch detected!  Assume -R? [n]
Apply anyway? [n]
2 out of 2 hunks ignored -- saving rejects to file drivers/usb/class/cdc-acm.c.rej
+ echo '*** Patch patches.fixes/usb-fix-oops-in-acm-disconnect.patch failed ***'
*** Patch patches.fixes/usb-fix-oops-in-acm-disconnect.patch failed ***
Comment 21 Greg Kroah-Hartman 2006-02-07 22:06:32 UTC
Ugh, you are correct, my appologies.  I had forgotten to clean out my patches
when Linus accepted them into mainline...
Comment 22 Greg Kroah-Hartman 2006-02-07 22:08:47 UTC
ok, this is related to the slab debugging stuff (well, it's showing it up
better.)  I thought this had been fixed :(

Can you mount usbfs in /proc/bus/usb/ and attach the output of the /proc/bus/usb/devices file?
Comment 23 Dirk Mueller 2006-02-07 22:12:24 UTC
Created attachment 66888 [details]
/proc/bus/usb/devices
Comment 24 Dirk Mueller 2006-02-07 22:20:19 UTC
bug 148872 might be related
Comment 25 Dirk Mueller 2006-02-15 10:09:22 UTC
Created attachment 68571 [details]
7th oops

still oopses
Comment 26 Greg Kroah-Hartman 2006-03-06 15:50:16 UTC
*** Bug 155396 has been marked as a duplicate of this bug. ***
Comment 27 Dirk Mueller 2006-03-06 15:57:12 UTC
needless to say, it still oopses. 
Comment 28 Greg Kroah-Hartman 2006-03-15 23:01:11 UTC
Created attachment 73122 [details]
Potential patch to fix the cdc-acm oops

Can you try this patch out to see if it fixes the 
issue for you?
Comment 29 Forgotten User ZhJd0F0L3x 2006-03-19 18:47:56 UTC
do i need to recompile cdc-acm, vmlinuz or everything to test that?
If "everything", then it would be nice if you could provide a test-kernel rpm :-)
Comment 30 Dirk Mueller 2006-03-21 19:45:00 UTC
yes, works fine now
Comment 31 Greg Kroah-Hartman 2006-04-12 16:42:00 UTC
This should be now fixed in the recent beta releases.  If not, please reopen.
Comment 32 Forgotten User ZhJd0F0L3x 2006-04-27 10:41:20 UTC
verified with 10.1RC1 - i can do all sorts of evil stuff to my usb connection and the kernel survives it :-)
Thanks!