[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 6 Jun 2008 16:03:35 +0300 (EEST)
From: "Ilpo Järvinen" <ilpo.jarvinen@...sinki.fi>
To: Alexander Beregalov <a.beregalov@...il.com>
cc: David Miller <davem@...emloft.net>, kernel-testers@...r.kernel.org,
Netdev <netdev@...r.kernel.org>
Subject: Re: WARNING: at net/ipv4/tcp_input.c:2213
On Thu, 5 Jun 2008, Alexander Beregalov wrote:
> 2008/6/5 David Miller <davem@...emloft.net>:
> > From: "Alexander Beregalov" <a.beregalov@...il.com>
> > Date: Thu, 5 Jun 2008 01:50:49 +0400
> >
> >> I have got the following message:
> >
> > We need to know the exact kernel version you were running, what
> > kinds of things the system was doing at the time, etc.
> >
>
> It is 2.6.26-rc3-00332-gcbff290.
> Host runs as a NFS server and rtorrent client.
> That is all network activity on it.
> I do not know what system was doing at this time, I have found this
> warning only after some time.
Thanks for reporting though I'm afraid that until there's somebody able
to reproduce it at least semi-often, this will remain unsolved. It's
probably another reno all-segments already sacked_out situation where then
one of them gets marked as lost in addition to that similar to the
tcp_simple_retransmit that is already fixed in the kernel you're using.
...In case you're worried, I'd say it has been there years and it won't
corrupt anything anyway and the situation resolved itself on some later
executed path (otherwise you'd see more than those two messages). They're
visible only because I recently (in 2.6.23 or .24) started tracking these
invariants by making them to complain and number of things have been
already fixed because of that :-).
It seems that there is a small number of reports @ kerneloops.org for
this same place but the count is noticably smaller (nearly by magnitude)
than what it has been with the same type issues that were finally fixed
couple of days ago, ie., after months of waiting. So there might not be
some near-term breakthrough coming up :-).
--
i.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists