race in fsnotify subsystem causes kernel oops
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Linux |
Fix Released
|
Medium
|
|||
linux (Ubuntu) |
Fix Released
|
High
|
Chris J Arges | ||
Precise |
Fix Released
|
High
|
Chris J Arges | ||
Quantal |
Fix Released
|
High
|
Chris J Arges | ||
Raring |
Fix Released
|
High
|
Chris J Arges |
Bug Description
This race can be triggered by plugging and unplugging a usb disk.
Related upstream bug: https:/
Previous bug: https:/
This was fixed in the above bug using patches that were not mainline yet. This bug aims to re-introduce the proper patches using the properly mainline patches.
--
SRU Justification:
Impact:
When plugging and unplugging a USB drive occasionally a race condition in the notify subsystem causes a kernel oops.
Fix:
A set of patches from upstream address this issue, this should already
be present in raring.
Here are the commit hashes:
6960b0d909cde5b
64c20d2a20fce29
d5a335b845792d2
e2a29943e9a2ee2
986ab09807ca945
6dfbd149946c22c
104d06f08ea5924
23e964c284ca0a7
986129520479d68
Testcase:
Comment #8 and #9 in the upstream bug: https:/
In addition I've tested using the LTP tests for inotfy and these run properly with the fix applied.
Changed in linux (Ubuntu Quantal): | |
assignee: | nobody → Chris J Arges (christopherarges) |
Changed in linux (Ubuntu Precise): | |
assignee: | nobody → Chris J Arges (christopherarges) |
importance: | Undecided → High |
Changed in linux (Ubuntu Quantal): | |
importance: | Undecided → High |
Changed in linux (Ubuntu Precise): | |
status: | New → In Progress |
Changed in linux (Ubuntu Quantal): | |
status: | New → In Progress |
Changed in linux: | |
importance: | Unknown → Medium |
status: | Unknown → Expired |
tags: | added: verification-done-quantal |
Changed in linux: | |
status: | Expired → Fix Released |
Changed in linux (Ubuntu Raring): | |
status: | In Progress → Fix Released |
Changed in linux (Ubuntu): | |
status: | In Progress → Fix Released |
I'm tagging this bug as verified for Precise: running the reproducer (https:/ /lkml.org/ lkml/2012/ 12/20/536) with the kernel in -updates causes an oops; running it with the -proposed kernel doesn't.