From: Thomas Graf Subject: Re: PROBLEM: IProute hangs after running traffic shaping scripts Date: Mon, 8 Nov 2004 19:33:00 +0100 Message-ID: <20041108183300.GF31969@postel.suug.ch> References: <418B4C7C.8000402@crocom.com.pl> <418EA032.7050507@trash.net> <418ECE85.9090203@trash.net> <20041108135431.GE31969@postel.suug.ch> <418F9AD0.1040701@trash.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Szymon Miotk , netdev@oss.sgi.com Return-path: To: Patrick McHardy Content-Disposition: inline In-Reply-To: <418F9AD0.1040701@trash.net> Sender: netdev-bounce@oss.sgi.com Errors-to: netdev-bounce@oss.sgi.com List-Id: netdev.vger.kernel.org * Patrick McHardy <418F9AD0.1040701@trash.net> 2004-11-08 17:12 > Thomas Graf wrote: > > >>+ /* unlink inner qdiscs from dev->qdisc_list immediately */ > >>+ if (qdisc->ops->cl_ops != NULL) > >>+ list_add(&qdisc->list, &cql); > >> > >> > > > >I think you should extend the above to: > > > >if (qdisc->ops->cl_ops != NULL && !list_empty(&qdisc->list)) > > > >Otherwise you might unlink entries just added before the > >rcu callback and it prevents an unneeded attempt to delete > >qdiscs for inner classful qdiscs > > > > > There is some optimization possible, I will do this for the final > patch. But I don't understand the problem you refer to, can you > please explain ? I don't have the time to verify this at the moment but: 1) qdisc_destroy unlinking all the lists 2) RTM_NEWQDISC creating a new qdisc with the same major classid as the old one which will suceed since the old one cannot be found anymore. 3) rcu callback __qdisc_destroy -> qdisc_destroy looking through qdisc_list again and then deleting the new entries because their major classid matches. I might be missing something though.