From mboxrd@z Thu Jan 1 00:00:00 1970 From: Patrick McHardy Subject: Re: again oops: conntrack & 2.6.14-git14 Date: Thu, 01 Dec 2005 11:40:39 +0100 Message-ID: <438ED327.8030509@trash.net> References: <43760B97.9010603@fliegl.de> <4380B5B1.9040506@trash.net> <4384A904.4080700@fliegl.de> <4388730A.2060003@trash.net> <438877CE.2050305@fliegl.de> <4388F1F7.5060402@trash.net> <438ED11C.4090508@fliegl.de> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: netfilter-devel@lists.netfilter.org Return-path: To: Deti Fliegl In-Reply-To: <438ED11C.4090508@fliegl.de> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: netfilter-devel-bounces@lists.netfilter.org Errors-To: netfilter-devel-bounces@lists.netfilter.org List-Id: netfilter-devel.vger.kernel.org Deti Fliegl wrote: > Patrick McHardy wrote: > >> The crashes seem to be unrelated to conntrack. They all happen in >> show_trace+489 because of accesses to 0xffffffff82800000, which >> is the end of the kernel text mapping on x86_64. Your stress test >> seems to trigger the softlockup watchdog, which is why you are >> seeing these. I suggest you resend your report to lkml. > > Then please explain to me why it does only happen when calling > 'conntrack -L'. The system is 100% stable, there is no defective > hardware and no weird configuration. Under same conditions I can run > kernel compilation jobs in a row... See above. It seems to trigger the softlockup watchdog for some reason, which calls show_trace, which is where the crash happens.