From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarek Poplawski Subject: Re: [Bugme-new] [Bug 19692] New: linux-2.6.36-rc5 crash with gianfar ethernet at full line rate traffic Date: Sat, 16 Oct 2010 21:48:15 +0200 Message-ID: <20101016194815.GA1894@del.dom.local> References: <20101010103236.GA1919@del.dom.local> <20101015085853.GA8091@ff.dom.local> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Andrew Morton , netdev@vger.kernel.org, bugzilla-daemon@bugzilla.kernel.org, bugme-daemon@bugzilla.kernel.org, Anton Vorontsov To: emin ak Return-path: Received: from mail-wy0-f174.google.com ([74.125.82.174]:39275 "EHLO mail-wy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754079Ab0JPTsV (ORCPT ); Sat, 16 Oct 2010 15:48:21 -0400 Received: by wyb28 with SMTP id 28so1764478wyb.19 for ; Sat, 16 Oct 2010 12:48:20 -0700 (PDT) Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Sat, Oct 16, 2010 at 02:14:01AM +0300, emin ak wrote: > Hi Jarek. > Sorry for delayed answer. As I promised, I have started the tests on > Monday (in spite of dealing with The Monday Syndrome:) I had applied > your patch and after two billions packet (and approximatly four hours) > passed kernel crashed with skb_over_panic error similar with first > type. To ensure the patch is failed or not, I rerun the same test > again. That time, surprisingly, it did'nt crashed again for two days > with same kernel. But this situation had occured before, I think > sometimes because of randomness of the applied ethernet traffic and > mostly because I cant apply all full line rate random traffic to my > target device because of wrong test setup (switch / hardware packet > generator settings etc..), it takes three or more day to crash the > kernel and sometimes it never crashes. My device only and only crashes > when I can apply full line rate random traffic. Before informing you > and the list with (official) test results, I want to be sure with the > truth of them. So that please let me apply more test to the target > device for a fews day more. After that I wish I'll came with good > news! Hi Emin, Sorry for being impatient. Actually, waiting is no problem for me. I simply didn't know how much this bug is reproducible, and was a bit mislead by your forecast of one day result (and terrified btw seeing words Monday and immediately side by side ;-) So, a few days or weeks, no problem, it's all up to you. On the other hand, it looks like there might be something else/more, so I'd suggest to add this last skb_over_panic to bugzilla too. Thanks, Jarek P.