From: Nick Piggin Subject: Re: [patch 6/6] mm: fsync livelock avoidance Date: Thu, 11 Dec 2008 23:45:14 +0100 Message-ID: <20081211224514.GE8294@wotan.suse.de> References: <20081210072454.GB27096@wotan.suse.de> <20081210074209.GG27096@wotan.suse.de> <20081211142347.2546b16c.akpm@linux-foundation.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-fsdevel@vger.kernel.org, mpatocka@redhat.com To: Andrew Morton Return-path: Received: from cantor2.suse.de ([195.135.220.15]:53845 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757314AbYLKWpQ (ORCPT ); Thu, 11 Dec 2008 17:45:16 -0500 Content-Disposition: inline In-Reply-To: <20081211142347.2546b16c.akpm@linux-foundation.org> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: On Thu, Dec 11, 2008 at 02:23:47PM -0800, Andrew Morton wrote: > obtw, > > On Wed, 10 Dec 2008 08:42:09 +0100 > Nick Piggin wrote: > > > For simplicity, I have removed the "don't wait for writeout if we hit -EIO" > > logic from a couple of places. I don't know if this is really worth the added > > complexity (EIO will still get reported, but it will just take a bit longer; > > an app can't rely in specific behaviour or timeliness here). > > This is ungood. The device layer likes to twiddle thumbs for 30 > seconds or more when it hits an IO error. We went and made that 30,000 > or more.. It isn't really a good solution anyway, because I think it's much less likely for writepage to return -EIO directly. Usually they would come back via data IO completion asynchronously. And if we are fsyncing so many requests anyway, we are likely going to start blocking behind them in the submission path anyway (elevator queues fill up).