Return-path: Received: from rhun.apana.org.au ([64.62.148.172]:40760 "EHLO arnor.apana.org.au" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1753157AbYGTOcp (ORCPT ); Sun, 20 Jul 2008 10:32:45 -0400 From: Herbert Xu To: davem@davemloft.net (David Miller) Subject: Re: [PATCH 20/31]: pkt_sched: Perform bulk of qdisc destruction in RCU. Cc: kaber@trash.net, hadi@cyberus.ca, netdev@vger.kernel.org, johannes@sipsolutions.net, linux-wireless@vger.kernel.org In-Reply-To: <20080717.152447.89672084.davem@davemloft.net> Message-Id: (sfid-20080720_163252_316320_77D2DBBB) Date: Sun, 20 Jul 2008 22:32:35 +0800 Sender: linux-wireless-owner@vger.kernel.org List-ID: David Miller wrote: > > Are we really really sure it's OK to let users do this? :) Well security questions aside, I have seen this used on routers to ensure certain traffic (e.g., VOIP) get preferential treatment when it comes to transmission. Of course some of these routers setup custom qdiscs so they wouldn't be affected, but there may be ones out there relying on the default pfifo. Now I understand that having a pfifo is an anathema to multi-queue TX by definition. However, even if we couldn't preserve the full semantics by default, if we could at least preserve the ordering within each queue it would still be a plus. Cheers, -- Visit Openswan at http://www.openswan.org/ Email: Herbert Xu ~{PmV>HI~} Home Page: http://gondor.apana.org.au/~herbert/ PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt