From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [iproute2] iproute2: Allow 'ip addr flush' to loop more than 10 times. Date: Tue, 29 Jun 2010 10:02:26 -0700 (PDT) Message-ID: <20100629.100226.235686154.davem@davemloft.net> References: <4C29925B.9090008@candelatech.com> <20100628.233600.242129599.davem@davemloft.net> <4C2A0CF3.9020204@candelatech.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: greearb@gmail.com, netdev@vger.kernel.org To: greearb@candelatech.com Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:38734 "EHLO sunset.davemloft.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932229Ab0F2RCN (ORCPT ); Tue, 29 Jun 2010 13:02:13 -0400 In-Reply-To: <4C2A0CF3.9020204@candelatech.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Ben Greear Date: Tue, 29 Jun 2010 08:10:43 -0700 > If I understand your proposal properly, this would seem to be > somewhat O(N^2) if we have large numbers of addresses, and I'm > hoping to support thousands of IPs with decent performance. I am not modifying the computational complexity or cost of the operation at all. I'm just changing under what circumstances it gives up. > What do you think about improving the kernel side so that we can send > a single netlink msg to delete all addresses on an interface, and just > let the kernel do the looping/locking needed to make it happen? I think the current scheme isn't that bad.