From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: Linux kernel patch: elide fib_validate_source() completely when possible - bad side effect? Date: Wed, 22 Jan 2014 18:53:05 -0800 (PST) Message-ID: <20140122.185305.1125981867482223830.davem@davemloft.net> References: <52DCF1440200005D000477FD@gwia2.atlnz.lc> <20140123020224.GG7269@order.stressinduktion.org> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: ja@ssi.bg, gregory.hoggarth@alliedtelesis.co.nz, netdev@vger.kernel.org To: hannes@stressinduktion.org Return-path: Received: from shards.monkeyblade.net ([149.20.54.216]:53036 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752394AbaAWCxJ (ORCPT ); Wed, 22 Jan 2014 21:53:09 -0500 In-Reply-To: <20140123020224.GG7269@order.stressinduktion.org> Sender: netdev-owner@vger.kernel.org List-ID: From: Hannes Frederic Sowa Date: Thu, 23 Jan 2014 03:02:24 +0100 > On Wed, Jan 22, 2014 at 11:04:52PM +0200, Julian Anastasov wrote: >> I don't remember someone mentioning about such >> side-effect, I guess it is overlooked. IMHO, it is not a good >> reason to restore the old behavior. Lets see other opinions. > > A routing lookup to check for broadcast is much too expansive, I agree. But > can't we just check skb->pkt_type == PACKET_BROADCAST? We can't universally know what a broadcast is, because the prefix of every network is a local piece of information. We might not have the route necessary to know the source address is a broadcast if the packet came through more than 1 hop already.