From: Thomas Graf <tgraf@infradead.org>
Date: Mon, 12 Apr 2010 04:39:03 -0400

quoted text
> On Mon, Apr 12, 2010 at 04:06:33AM -0400, Thomas Graf wrote:
>>  - While the issue is appearing, the acceptq seems to be overflowing. Both
>>    LISTENOVERFLOWS and LISTENDROPS are increasing although not by the exact
>>    number of delay occurences. inetdiag reports sk_max_ack_backlog to be 0
>>    therefore one possibility that comes to mind is that sk_ack_backlog
>>    underflows due to a race.
> 
> Forget about the underflow thought, inetdiag was reporting falsely.
> sk_max_ack_backlog is set to 128 as it should and the listen overflow
> happens normally. Still the fact remains that while the issue is appearing
> listen overflows are counted.

I can't reproduce on my system even with sched_child_runs_first set to '1'.

Are you running identd or something like that which intercepts the connections
to port 22 before 'sshd' actually gets it?
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html