Bug 2073 – NDisc cache entries update timer might be stuck in a loop

Bug 2073 - NDisc cache entries update timer might be stuck in a loop


Summary:	NDisc cache entries update timer might be stuck in a loop

Status:	RESOLVED FIXED

Product:	ns-3
Classification:	Unclassified
Component:	ipv6
Version:	ns-3-dev
Hardware:	All All

Importance:	P5 normal
Assigned To:	Tommaso Pecorella

URL:
Keywords:

Depends on:
Blocks:
	Show dependency tree / graph

Reported:	2015-02-27 03:44 UTC by Tommaso Pecorella
Modified:	2015-02-28 09:20 UTC (History)
CC List:	1 user (show)

See Also:

Attachments
test program (1.84 KB, text/x-csrc) 2015-02-27 03:44 UTC, Tommaso Pecorella	Details
patch (10.54 KB, patch) 2015-02-28 09:05 UTC, Tommaso Pecorella	Details \| Diff
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.

Description Tommaso Pecorella 2015-02-27 03:44:56 UTC

Created attachment 1980 [details]
test program

The effect is summarized in the simple example attached. One of the nodes keeps probing the other one, even after receiving the correct reply.

The bug was reported by Julien Vaubourg here:
https://groups.google.com/forum/#!topic/ns-3-users/t_SMjHCjGh0

The problem is in the NDisc cache timers. Each state have its own timer and it's not easy to make sure that the other state timers are expired when a transition occurs.
The solution is a partial code refactoring, in order to simplifying the code and fix the error.
There should be only one timer, and it should be started/stopped by the state transition.

Comment 1 Tommaso Pecorella 2015-02-28 09:05:56 UTC

Created attachment 1981 [details]
patch

The patch simplifies the NUD timers (we just need one).
Of course it also removes the bug, as it was due to a timer not being correctly stopped.

Comment 2 Tommaso Pecorella 2015-02-28 09:20:17 UTC

Pushed in changeset 11225:750bd0fc8706