From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Denis V. Lunev" Subject: Re: [PATCH] remove __net_initdata attribute Date: Fri, 26 Oct 2007 15:28:32 +0400 Message-ID: <4721CF60.5060600@sw.ru> References: <20071026095652.GA8788@iris.sw.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: davem@davemloft.net, devel@openvz.org, netdev@vger.kernel.org, containers@lists.osdl.org, benjamin.thery@bull.net, clg@fr.ibm.com To: "Denis V. Lunev" Return-path: Received: from swsoft-mipt-nat.sw.ru ([195.214.233.10]:53354 "EHLO iris" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1750905AbXJZLZo (ORCPT ); Fri, 26 Oct 2007 07:25:44 -0400 In-Reply-To: <20071026095652.GA8788@iris.sw.ru> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Dave, I am sorry, plz disregard the patch :( - if the file is a part of the kernel, __exit (and pernet_unregister) will never called and should be dropped - if the file is a part of a module, __init is NOT dropped on the module loading stage So, we can still make embedded people happy :) Thanks Pavel to pointing me this out. Denis V. Lunev wrote: > The discussion with Eric W. Biederman reveals that the concept for > __net_initdata is bogus. The ->exit method of pernet_operations > should be called during module exit. > > So, discard this attribute. > > Signed-off-by: Denis V. Lunev