Debian 6 Squeeze bug report: application or kernel bug?
I am not sure where to file this bug report, because it's a kerneloops caused by a bug.
This is what happened:
1. USB2 hard disk disconnected and immediately reconnected, which could be caused by a known bug in certain EHCI controller chips
2. An X process probably got terminated and restarted
3. Checking /etc/mtab, I noticed the two partitions on the old device path are still registered as mounted
4. An attempt to unmount either one of them fails, at which point another stack trace is being printed - so yet another severe (unhandled) error or exception occurred
5. I notice that several processes are in a blocked state
6. The blocked processes prevent a normal shutdown from the desktop
7. Terminating the blocked processes fails (using 'kill -KILL #PID' and 'kill -TERM #PID'), in any possible way this command didn't work
8. As a last resort, I switched to a basic terminal trying to shutdown using 'init 6', which failed either
9. Decided to take the chance and use the case's reset button
*******************************
* KEY LINES IN 'syslog' *
*******************************
kernel: [ 1406.865500] usb 1-3: USB disconnect, address 2
kernel: [ 1407.144039] usb 1-3: new high speed USB device using ehci_hcd and address 3
(...)
kernel: [ 1407.281417] scsi6 : SCSI emulation for USB Mass Storage devices
(... so far, nothing unusual for a known bug, but from the third line below it is unusual ...)
kernel: [ 1412.339510] sd 6:0:0:0: [sde] Attached SCSI disk
kernel: [ 1412.873626] EXT4-fs error (device sdd2): ext4_find_entry: reading directory #2 offset 0
kernel: [ 1412.873695] BUG: unable to handle kernel paging request at 015ef000
kernel: [ 1412.873706] IP: [<c1140c27>] __percpu_counter_add+0x21/0x6d
kernel: [ 1412.873727] *pde = 00000000
kernel: [ 1412.873735] Oops: 0000 [#1] SMP
(... next occurrance: ...)
kernel: [ 1647.035866] BUG: unable to handle kernel paging request at 015ef000
kernel: [ 1647.035878] IP: [<c1140c27>] __percpu_counter_add+0x21/0x6d
kernel: [ 1647.035895] *pde = 00000000
kernel: [ 1647.035901] Oops: 0000 [#2] SMP
kernel: [ 1647.035914] Modules linked in: usb_storage speedstep_lib cpufreq_stats cpufreq_userspace cpufreq_conservative cpufreq_powersave fuse ext2 (and many more) [last unloaded: scsi_wait_scan]
(...)
kernel: [ 1647.036063] Pid: 3218, comm: umount Tainted: G D (2.6.32-5-686 #1)
and onto the register contents, stacktrace etc. again
It concerns me that things got this far now. I encountered the problem at #1 more often while using the MEPIS Live CD, but it simply resulted in read error messages, but nothing serious. This time, at least one of the exceptions occurred when 'umount' was running.
Should I report this as a kernel bug, or as a bug in the mount package? (considering "Pid: 3218, comm: umount ...")
Reply to: