Discussion

• Extensive discussion of bug which allowed mbuf leaks on RACK when packets didn't contain timestamps. There appears to be a mystery where removing an unrelated patch (related to locking and timing of upcalls) reduces the severity of the leak. rrs@ will try to recreate the mysterious behavior so we can understand it better.

• PR 256657: tuexen@ cannot reproduce the problem. It is hard to understand why the problem is not reproducible.

• D30740: Discussed whether we should be specific about vendors which are non-compliant. We decided to make a statement in the man page about the danger of disabling the sysctl, but not name specific vendor names.

• D29441: Unit tested. rrs@ approved.

• Will skip next meeting. Next meeting is July 15.

Next meeting: 15 July @ 1400 UTC