[Bug 551948] New: kernel BUG at /usr/src/packages/BUILD/kernel-desktop-2.6.31.3/linux-2.6.31/fs/dcache.c:669!
http://bugzilla.novell.com/show_bug.cgi?id=551948 Summary: kernel BUG at /usr/src/packages/BUILD/kernel-desktop-2.6.31.3/linux- 2.6.31/fs/dcache.c:669! Classification: openSUSE Product: openSUSE 11.2 Version: RC 2 Platform: i586 OS/Version: openSUSE 11.2 Status: NEW Severity: Normal Priority: P5 - None Component: Kernel AssignedTo: kernel-maintainers@forge.provo.novell.com ReportedBy: mawa@uni-freiburg.de QAContact: qa@suse.de Found By: --- User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1.4) Gecko/20091016 SUSE/3.5.4-1.1.2 Firefox/3.5.4 the system reproducable gets a kernel bug. perhaps a problem with nfs4 mounts? occured in 11.2RC1 and RC2 Oct 19 00:25:36 localhost kernel: [22216.254786] kernel BUG at /usr/src/packages/BUILD/kernel-desktop-2.6.31.3/linux-2.6.31/fs/dcache.c:669! Nov 2 00:05:32 localhost kernel: [43543.498644] kernel BUG at /usr/src/packages/BUILD/kernel-desktop-2.6.31.5/linux-2.6.31/fs/dcache.c:669! for full logs see attachment. Reproducible: Always Steps to Reproduce: 1. 2. 3. Expected Results: no kernel bug ;) -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 User mawa@uni-freiburg.de added comment http://bugzilla.novell.com/show_bug.cgi?id=551948#c1 --- Comment #1 from Martin Walter <mawa@uni-freiburg.de> 2009-11-02 12:55:39 MST --- Created an attachment (id=325182) --> (http://bugzilla.novell.com/attachment.cgi?id=325182) logs -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c Jeff Mahoney <jeffm@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P5 - None |P2 - High CC| |nfbrown@novell.com AssignedTo|kernel-maintainers@forge.pr |sjayaraman@novell.com |ovo.novell.com | -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c2 Suresh Jayaraman <sjayaraman@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED --- Comment #2 from Suresh Jayaraman <sjayaraman@novell.com> 2009-12-07 10:45:56 UTC --- Thanks for the bug report. I'll take look. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c3 Martin Walter <mawa@uni-freiburg.de> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |mawa@uni-freiburg.de --- Comment #3 from Martin Walter <mawa@uni-freiburg.de> 2009-12-07 11:35:57 UTC --- Thanx. The problem still exists under 11.2 final. It appears during a shutdown when performing a "rcnfsserver stop". But only, if I had an NFS4 (with krb5p) mount before, with some traffic to/from the NFS4 directory. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c4 Suresh Jayaraman <sjayaraman@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |mawa@uni-freiburg.de --- Comment #4 from Suresh Jayaraman <sjayaraman@novell.com> 2009-12-08 13:58:04 UTC --- Comment #3 was very useful, Thanks. This looks like a dentry reference counting problem reproducible only while using krb5/i/p modes (i.e. /var/lib/nfs/rpc_pipefs used by rpc.gssd for upcalls from kernel). However, I have not figured out where exactly the problem is. Could you please try enabling additional debugging information by doing the following? Enable debug just before unmounting #dmesg -c (clears kernel ring buffer) #rpcdebug -m nfs vfs dircache lookupcache file client mount #umount -alt nfs /mnt/point (umount all nfs mounts lazily) #dmesg > nfs-debug.txt attach nfs-debug.txt -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c5 --- Comment #5 from Martin Walter <mawa@uni-freiburg.de> 2009-12-08 19:08:27 UTC --- Created an attachment (id=331613) --> (http://bugzilla.novell.com/attachment.cgi?id=331613) output of "dmesg" the problem occured not after "umount -alt nfs" but "rcnfs stop". -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c6 --- Comment #6 from Suresh Jayaraman <sjayaraman@novell.com> 2009-12-09 12:29:04 UTC --- Created an attachment (id=331725) --> (http://bugzilla.novell.com/attachment.cgi?id=331725) Proposed patch The previous log didn't help much because what could help more is enabling rpc debug messages than nfs debugging. However, I think I see the problem with the rpc_pipefs code but I'm not too sure since I don't have the setup. Would it mind giving this patch a try? Also this time please enable rpc debugging: #dmesg -c (clears kernel ring buffer) #rpcdebug -m rpc auth cache #rcnfs stop #dmesg > rpc-debug.txt -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c7 --- Comment #7 from Martin Walter <mawa@uni-freiburg.de> 2009-12-09 23:18:06 UTC --- Created an attachment (id=331860) --> (http://bugzilla.novell.com/attachment.cgi?id=331860) "dmesg" after "rpcdebug -m rpc auth cache; rcnfs stop" -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c8 --- Comment #8 from Suresh Jayaraman <sjayaraman@novell.com> 2010-01-05 10:25:21 UTC --- Could you please try the patch attached in Comment #6 and see whether the problem goes away while I'm working on other priority items? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c9 --- Comment #9 from Martin Walter <mawa@uni-freiburg.de> 2010-01-05 12:14:11 UTC --- yes. but I have no build environment nor experience. so could you make me a kernel rpm? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c10 --- Comment #10 from Suresh Jayaraman <sjayaraman@novell.com> 2010-01-06 11:40:12 UTC --- I have copied the test kernel built with fix here: http://www.suse.com/~sjayaraman/testkernels/ (it might a few hours to sync) Note that this is a test kernel and not recommended for use in production machines. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c11 --- Comment #11 from Martin Walter <mawa@uni-freiburg.de> 2010-01-06 17:24:42 UTC --- all ok now! :-) thank you very much! -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c12 --- Comment #12 from Suresh Jayaraman <sjayaraman@novell.com> 2010-01-08 06:12:34 UTC --- Good, however I don't see the need for that patch unless you have auth_unix/auth_sys mounts together with krb5 mounts on the same client. Is this the case? If the error reproducible when you have only krb5 mounts on the client? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c13 --- Comment #13 from Suresh Jayaraman <sjayaraman@novell.com> 2010-01-08 06:27:50 UTC --- Also please provide output of cat /proc/mounts. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c14 --- Comment #14 from Martin Walter <mawa@uni-freiburg.de> 2010-01-09 12:13:16 UTC --- Yes, I need both. We use diskless workstations and loginserver. --> http://www.rz.uni-freiburg.de/services/pc/netpoint --> http://www.rz.uni-freiburg.de/services/server/loginserver A typical /proc/mounts is attached in proc_mounts.txt . -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c15 --- Comment #15 from Martin Walter <mawa@uni-freiburg.de> 2010-01-09 12:15:25 UTC --- Created an attachment (id=335772) --> (http://bugzilla.novell.com/attachment.cgi?id=335772) cat /proc/mounts -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c16 --- Comment #16 from Suresh Jayaraman <sjayaraman@novell.com> 2010-01-11 13:23:27 UTC --- Hmm.. odd, I tried to reproduce this with 11.2 (server and client both) and I could not reproduce this issue. Is this reproducible with 11.2 as server? In you case I think the server is NetApp filer. If only reproducible with NetApp server, could you capture the network packets during unmount by doing: tcpdump -s0 -wnfs-debug.cap port 2049 (on the client) -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c17 --- Comment #17 from Martin Walter <mawa@uni-freiburg.de> 2010-01-11 14:17:14 UTC --- sorry for the confusion. the bug occured at home where i have 11.2 server and client with kernel 2.6.31.5-0.1. the /proc/mounts listing was from our university diskless project, which runs under 11.1 and which i can not upgrade to 11.2 because of this bug. but in my home environment your patch solved the bug, so i would suggest to just include your patch in the next kernel-update. in my tests the bug occured not always after only short work. but it occured always during shutdown after working a longer time (15 min?). -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c18 --- Comment #18 from Suresh Jayaraman <sjayaraman@novell.com> 2010-01-12 06:13:16 UTC --- My patch is just a band-aid even if it fixed the kernel BUG and not it doesn't fix the real problem. The real problem might lead to leaks even with the above patch. In my reproduction attempts I could never hit the problem and the code in question i.e. when rpc_client is destroyed seems correct. I'm discussing this issue upstream as well. What would help is a simple test case that can reproduce the issue everytime. It's not clear from Comment #3 whether the bug is noticed on the server or the client, do you see the error on the client? or both the server and the client are the same (loopback mounts)? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c19 --- Comment #19 from Martin Walter <mawa@uni-freiburg.de> 2010-01-12 10:50:16 UTC --- sorry, rcnfsserver is false in comment #3. it should read rcnfs as in comment #5 and #6. i try to find an easy way to reproduce the bug. perhaps it has to do with some programs i use (icewm, firefox, mutt, xterm) or kind of use (locking, pipe, socket) ? -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c20 --- Comment #20 from Suresh Jayaraman <sjayaraman@novell.com> 2010-01-28 05:14:12 UTC --- I came across a similar BUG recently which involves DM/LVM. Wondering whether you are using any of these.. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c21 Martin Walter <mawa@uni-freiburg.de> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|mawa@uni-freiburg.de | --- Comment #21 from Martin Walter <mawa@uni-freiburg.de> 2010-01-30 16:02:36 UTC --- Yes, I have LVM in use. Today I made a yast-online-update and moved from your kernel-desktop-2.6.31.9-1.i586 to kernel-desktop-2.6.31.8-0.1.1.i586 from YOU. Then I tried hard to reproduce the bug but without success. So it seems the bug is gone! :-) -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=551948 http://bugzilla.novell.com/show_bug.cgi?id=551948#c22 Suresh Jayaraman <sjayaraman@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #22 from Suresh Jayaraman <sjayaraman@novell.com> 2010-02-01 18:21:02 UTC --- Hmm, not clear what made the bug go away. Marking it as RESOLVED FIXED. In case if the bug reappears please reopen this bug with possibly a reproducer. Thanks! -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com