From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: panic 2.6.27-rc3-git2, qdisc_dequeue_head Date: Sun, 17 Aug 2008 15:02:58 -0700 (PDT) Message-ID: <20080817.150258.117813919.davem@davemloft.net> References: <20080817133608.GA3584@ami.dom.local> <200808172236.17857.denys@visp.net.lb> <20080817214752.GA2546@ami.dom.local> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: denys@visp.net.lb, netdev@vger.kernel.org To: jarkao2@gmail.com Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:52508 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751546AbYHQWC7 (ORCPT ); Sun, 17 Aug 2008 18:02:59 -0400 In-Reply-To: <20080817214752.GA2546@ami.dom.local> Sender: netdev-owner@vger.kernel.org List-ID: From: Jarek Poplawski Date: Sun, 17 Aug 2008 23:47:52 +0200 > @@ -638,9 +638,9 @@ static void notify_and_destroy(struct sk_buff *skb, struct nlmsghdr *n, u32 clid > qdisc_notify(skb, n, clid, old, new); > > if (old) { > - spin_lock_bh(&old->q.lock); > + sch_tree_lock(old); > qdisc_destroy(old); > - spin_unlock_bh(&old->q.lock); > + sch_tree_unlock(old); > } > } > Ok, this makes sense, for the classful qdisc case. If we really wanted to be super-careful, we could remember the root qdisc at the top of qdisc_graft() and pass that into notify_and_destroy().