Return-Path: linux-nfs-owner@vger.kernel.org Received: from mail-qa0-f52.google.com ([209.85.216.52]:35346 "EHLO mail-qa0-f52.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755207Ab3BAAIX (ORCPT ); Thu, 31 Jan 2013 19:08:23 -0500 Date: Thu, 31 Jan 2013 16:08:12 -0800 From: Tejun Heo To: "Myklebust, Trond" Cc: "J. Bruce Fields" , "akpm@linux-foundation.org" , "rusty@rustcorp.com.au" , "linux-kernel@vger.kernel.org" , "linux-nfs@vger.kernel.org" Subject: Re: [PATCH 11/14] nfs: idr_destroy() no longer needs idr_remove_all() Message-ID: <20130201000812.GO6824@mtj.dyndns.org> References: <1359163872-1949-1-git-send-email-tj@kernel.org> <1359163872-1949-12-git-send-email-tj@kernel.org> <20130129225848.GD6219@fieldses.org> <4FA345DA4F4AE44899BD2B03EEEC2FA9183426FE@sacexcmbx05-prd.hq.netapp.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <4FA345DA4F4AE44899BD2B03EEEC2FA9183426FE@sacexcmbx05-prd.hq.netapp.com> Sender: linux-nfs-owner@vger.kernel.org List-ID: Hello, Trond. On Wed, Jan 30, 2013 at 02:18:06AM +0000, Myklebust, Trond wrote: > My worry is more about Tejun's comment that we did actually need a call > to idr_remove_all() in the original code. Do we need to queue up a fix > for the 3.8 and existing stable kernels? IIUC, the only case where idr_layer is left around after idr_remove() on all elems is when the deletions lead to multi-level left-most collapse. Not sure whether the current code can actually hit that. No harm in adding idr_remove_all(), I guess. Thanks. -- tejun