[Cluster-devel] [PATCH] idr: fix a subtle bug in idr_get_next()

[Hugh Dickins <hughd@google.com>]

On Sat, 2 Feb 2013, Randy Dunlap wrote:
> On 02/02/13 15:11, Tejun Heo wrote:
> > On Sat, Feb 02, 2013 at 03:10:48PM -0800, Tejun Heo wrote:
> >> Fix it by ensuring proceeding to the next slot doesn't carry over the
> >> unaligned offset - ie. use round_up(id + 1, slot_distance) instead of
> >> id += slot_distance.
> >>
> >> Signed-off-by: Tejun Heo <tj@kernel.org>
> >> Reported-by: David Teigland <teigland@redhat.com>
> >> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
> > 
> > David, can you please test whether the patch makes the skipped
> > deletion bug go away?
> > 
> > Thanks!
> 
> Hugh, did you update the idr test suite or the radix-tree test suite?

Sorry, not the answer you want: it was the radix-tree test harness, rtth,
that I updated a year ago (but then lib/radix-tree.c changed again).

Hugh

> 
> Is there an idr test suite or module?  I have a kernel module source file
> named idr_test.c by Eric Paris.
> 
> thanks,
> 
> -- 
> ~Randy