Return-Path: linux-nfs-owner@vger.kernel.org Received: from mout.perfora.net ([74.208.4.194]:56113 "EHLO mout.perfora.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S935338Ab3DHMGd (ORCPT ); Mon, 8 Apr 2013 08:06:33 -0400 Date: Mon, 8 Apr 2013 08:06:25 -0400 From: Jim Rees To: William Dauchy Cc: "Myklebust, Trond" , Linux NFS mailing list , Chuck Lever , "Schumaker, Bryan" Subject: Re: [PATCH 0/2] Stable patches for NFSv4/4.1 trunking Message-ID: <20130408120625.GA9924@umich.edu> References: <1365196588-25403-1-git-send-email-Trond.Myklebust@netapp.com> <1365280195.3651.4.camel@leira.trondhjem.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: Sender: linux-nfs-owner@vger.kernel.org List-ID: William Dauchy wrote: On Sat, Apr 6, 2013 at 10:29 PM, Myklebust, Trond wrote: > The main reason is that it should be a _very_ rare event, since it > requires a call to try_to_get_module(THIS) to fail, and so I can't see > that it could ever cause a huge leakage. > If someone can show that it is more of a problem than I suggest above, > then I'm happy to reconsider. Understood. I didn't know a patch fixing a very rare event could not go in stable. Every patch, no matter how "obviously" correct, carries a risk of introducing new bugs. Trond is responsible for weighing the risk of new bugs against the benefit of fixing old ones. He has a lot of experience doing this. If he makes a mistake, he takes the heat for your bugs. Introducing a new bug into stable is a lot worse than introducing one into -next.