From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jarek Poplawski Subject: Re: [BUG] NULL pointer dereference in skb_dequeue Date: Sun, 3 Aug 2008 12:08:50 +0200 Message-ID: <20080803100850.GA3077@ami.dom.local> References: <20080802.121815.167998262.davem@davemloft.net> <20080802201944.GA14983@ami.dom.local> <20080803092926.GA2971@ami.dom.local> <20080803.025616.218557268.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: emil.s.tantilov@intel.com, jeffrey.t.kirsher@intel.com, netdev@vger.kernel.org To: David Miller Return-path: Received: from ug-out-1314.google.com ([66.249.92.171]:10961 "EHLO ug-out-1314.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751787AbYHCKIT (ORCPT ); Sun, 3 Aug 2008 06:08:19 -0400 Received: by ug-out-1314.google.com with SMTP id h2so619367ugf.16 for ; Sun, 03 Aug 2008 03:08:18 -0700 (PDT) Content-Disposition: inline In-Reply-To: <20080803.025616.218557268.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: On Sun, Aug 03, 2008 at 02:56:16AM -0700, David Miller wrote: > From: Jarek Poplawski > Date: Sun, 3 Aug 2008 11:29:26 +0200 > > > After some re-checking one more question: why do you think this > > qdisc_root_lock() is safe as sch_tree_lock() (or anywhere else)? It > > seems, eg. during deactivation it can get root_lock of qdisc_default, > > and proceed with another qdisc? > > We hold RTNL at the time. > > If it's the default qdisc, that's fine, we'll reset it and free > up the packets when the RCU handler of the qdisc_destroy() runs. > > That's a case where locking the wrong qdisc is OK. Right! Thanks again, Jarek P.