From mboxrd@z Thu Jan 1 00:00:00 1970 From: Patrick McHardy Subject: Re: [RFC]: net-sched 00/05: dynamically sized class hashes Date: Wed, 02 Jul 2008 16:30:09 +0200 Message-ID: <486B90F1.8020606@trash.net> References: <20080701143410.26309.45560.sendpatchset@localhost.localdomain> <20080701.195050.177373433.davem@davemloft.net> <486B262F.2040108@cdi.cz> <486B54B7.2060100@trash.net> <486B8ECF.3070604@trash.net> <486B9064.4010506@cdi.cz> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: 7bit Cc: David Miller , netdev@vger.kernel.org To: Martin Devera Return-path: Received: from stinky.trash.net ([213.144.137.162]:39378 "EHLO stinky.trash.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753627AbYGBOaL (ORCPT ); Wed, 2 Jul 2008 10:30:11 -0400 In-Reply-To: <486B9064.4010506@cdi.cz> Sender: netdev-owner@vger.kernel.org List-ID: Martin Devera wrote: > Patrick McHardy wrote: >> Patrick McHardy wrote: >>> Martin Devera wrote: >>>> Interestingly, once you do this patch then children/siblings list-heads >>>> are used only in htb_parent_last_child and to detect empty class. >>>> Thus we could remove children/siblings lists altogether ane keep only >>>> children counter. >>>> I can't believe it is so simple, what do you think, Patrick ? >>> >>> Yes, it looks like that would work. I'll give it a try :) >> >> How about this, on top of my previous patches? > > Yes, it is exactly what have been in my mind. It should work. Thanks! > Only I'm wondering why I didn't do it this way in the first > version :-) I'm guessing there was no working example of destroying filters without walking the tree in hierarchical order. All other qdiscs were buggy or non-existant at that time :)