From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH 3/4] xfrm: remove policy lock when accessing policy->walk.dead Date: Wed, 31 Mar 2010 13:52:39 -0700 (PDT) Message-ID: <20100331.135239.170044408.davem@davemloft.net> References: <20100331131131.GA13793@gondor.apana.org.au> <1270042092.26743.26.camel@bigi> <20100331135331.GB14082@gondor.apana.org.au> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: hadi@cyberus.ca, timo.teras@iki.fi, netdev@vger.kernel.org To: herbert@gondor.apana.org.au Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:41120 "EHLO sunset.davemloft.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932107Ab0CaUwk (ORCPT ); Wed, 31 Mar 2010 16:52:40 -0400 In-Reply-To: <20100331135331.GB14082@gondor.apana.org.au> Sender: netdev-owner@vger.kernel.org List-ID: From: Herbert Xu Date: Wed, 31 Mar 2010 21:53:31 +0800 > Dave, please revert this patch. Herbert I actually don't agree with you on this one, sorry :-) If no semantic change occurred to the policy or state databases, there is zero reason to report the flush event, it's noise. Like Jamal, I too am pretty frustrated with how noisy and unusable the netlink messages are to watch when trying to debug something and this behavior change should break no applications whatsoever. When you have evidence of a real existing case failing because of Jamal's changes (not some case you crafted specifically to show that an empty flush doesn't get reported), I'll consider a revert or ask Jamal to fix that case up. Until then, Jamal's change is staying in the tree.