From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH 1/2] net: Let inet_diag and friends autoload Date: Wed, 17 Oct 2007 19:51:25 -0700 (PDT) Message-ID: <20071017.195125.15264659.davem@davemloft.net> References: <200710161728.23733.jdelvare@suse.de> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, kuznet@ms2.inr.ac.ru To: jdelvare@suse.de Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:60692 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1760204AbXJRCv0 (ORCPT ); Wed, 17 Oct 2007 22:51:26 -0400 In-Reply-To: <200710161728.23733.jdelvare@suse.de> Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org From: Jean Delvare Date: Tue, 16 Oct 2007 17:28:23 +0200 > By adding module aliases to inet_diag, tcp_diag and dccp_diag, we let > them load automatically as needed. This makes tools like "ss" run > faster. > > Signed-off-by: Jean Delvare > Cc: Alexey Kuznetsov > --- > The alias naming scheme for tcp_diag and dccp_diag follows what was > done for dccp_ipv4 and dccp_ipv6. I'm concerned it might be dangerous to load the module in this context, the RTNL semaphore is held and the module we are requesting could easily try to take it and deadlock. Currently we are lucky and this won't happen, but if the sub-proto registration is every changed to do so, it will be a nightmare for whoever is so unlucky to try and debug it.