From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: with ECARTIS (v1.0.0; list xfs); Thu, 09 Oct 2008 15:46:59 -0700 (PDT) Received: from cuda.sgi.com (cuda3.sgi.com [192.48.176.15]) by oss.sgi.com (8.12.11.20060308/8.12.11/SuSE Linux 0.7) with ESMTP id m99MkuX7030738 for ; Thu, 9 Oct 2008 15:46:56 -0700 Received: from ipmail05.adl2.internode.on.net (localhost [127.0.0.1]) by cuda.sgi.com (Spam Firewall) with ESMTP id 0574013D6FEB for ; Thu, 9 Oct 2008 15:48:35 -0700 (PDT) Received: from ipmail05.adl2.internode.on.net (ipmail05.adl2.internode.on.net [203.16.214.145]) by cuda.sgi.com with ESMTP id gAgIfID5Hyts0pYl for ; Thu, 09 Oct 2008 15:48:35 -0700 (PDT) Date: Fri, 10 Oct 2008 09:48:32 +1100 From: Dave Chinner Subject: Re: [PATCH V2] Re-dirty pages on ENOSPC when converting delayed allocations Message-ID: <20081009224832.GK9597@disturbed> References: <48EB1ABD.3020503@sgi.com> <20081009122726.GH9597@disturbed> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <20081009122726.GH9597@disturbed> Sender: xfs-bounce@oss.sgi.com Errors-to: xfs-bounce@oss.sgi.com List-Id: xfs To: Lachlan McIlroy , xfs-oss , xfs-dev On Thu, Oct 09, 2008 at 11:27:26PM +1100, Dave Chinner wrote: > On Tue, Oct 07, 2008 at 06:15:57PM +1000, Lachlan McIlroy wrote: > > If we get an error in xfs_page_state_convert() - and it's not EAGAIN - then > > we throw away the dirty page without converting the delayed allocation. This > > leaves delayed allocations that can never be removed and confuses code that > > expects a flush of the file to clear them. We need to re-dirty the page on > > error so we can try again later or report that the flush failed. > > Actually, those delalloc pages can be removed - they just need to > be handled in ->releasepage. The problem there is that the > delalloc state is checked by looking at the bufferhead, and by > the time we get to ->releasepage the buffer heads have already gone > through discard_buffer() and lost the buffer_delay() flag. > > IIRC I had a patch that did the delalloc conversion correctly in > ->releasepage by utilising a custom ->invalidatepage callouut, but > the performance overhead was very bad because it is done a page at a > time. ISTR even posting it to oss.... FWIW, here's the discussion: http://oss.sgi.com/archives/xfs/2008-04/msg00542.html I'm not saying that we Ń•hould be fixing this right now, but this is the underlying problem that is causing the direct I/O read to trip over stale delalloc extents once they've been invalidated. Cheers, Dave. -- Dave Chinner david@fromorbit.com