From: Peter Zijlstra Subject: Re: [PATCH] percpu_counter: use local_t and atomic_long_t if possible Date: Tue, 23 Dec 2008 12:43:21 +0100 Message-ID: <1230032601.9487.247.camel@twins> References: <4936D287.6090206@cosmosbay.com> <20081209214921.b3944687.akpm@linux-foundation.org> <49404925.7090902@cosmosbay.com> <200812121847.06432.rusty@rustcorp.com.au> <49424637.3010107@cosmosbay.com> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Cc: Andrew Morton , Rusty Russell , Theodore Tso , linux kernel , "David S. Miller" , Mingming Cao , linux-ext4@vger.kernel.org, Christoph Lameter To: Eric Dumazet Return-path: Received: from viefep19-int.chello.at ([62.179.121.39]:60892 "EHLO viefep18-int.chello.at" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1750912AbYLWLnY (ORCPT ); Tue, 23 Dec 2008 06:43:24 -0500 In-Reply-To: <49424637.3010107@cosmosbay.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: One think that sprung to mind,.. IFF we're going to implement percpu_counter with local_t and make local_t this funny tri-counter thing which has its own error, you need to fix up bdi_stat_error() - it expects to be an upper bound for the counter error, getting that wrong _will_ cause deadlocks.