From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarek Poplawski Subject: Re: panic 2.6.27-rc3-git2, qdisc_dequeue_head Date: Sun, 17 Aug 2008 12:31:41 +0200 Message-ID: <20080817103141.GC2907@ami.dom.local> References: <20080815190905.M56388@visp.net.lb> <200808171235.46413.denys@visp.net.lb> <20080817095220.GB2907@ami.dom.local> <200808171301.56002.denys@visp.net.lb> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: netdev@vger.kernel.org To: Denys Fedoryshchenko Return-path: Received: from nf-out-0910.google.com ([64.233.182.184]:32692 "EHLO nf-out-0910.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752323AbYHQKbd (ORCPT ); Sun, 17 Aug 2008 06:31:33 -0400 Received: by nf-out-0910.google.com with SMTP id d3so1004426nfc.21 for ; Sun, 17 Aug 2008 03:31:32 -0700 (PDT) Content-Disposition: inline In-Reply-To: <200808171301.56002.denys@visp.net.lb> Sender: netdev-owner@vger.kernel.org List-ID: On Sun, Aug 17, 2008 at 01:01:55PM +0300, Denys Fedoryshchenko wrote: > Ok, now things finally organized. > fixing patch #1 was missing. > > Rebooting system to this kernel. But strange thing, without fixing patch #1 it > didn't crash for whole night, and 1-3 hours now (with same options as > before). The fixing patch #1 was intended to fix only one issue shown in the warnings: qdisc_watchdog scheduled leaf qdiscs for qdisc_run(), so they were dequeued directly (skipping root qdiscs) and with wrong locks. The fixing patch #2 changes qdisc locking generally, so it probably fixed other locking issues btw. like this last oops seen after the patch #1. But still there is this (not critical) issue of skipping root qdiscs while dequeuing after qdisc_watchdog, which has to be done by the fixing patch #1. Jarek P.