From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from cuda.sgi.com (cuda1.sgi.com [192.48.157.11]) by oss.sgi.com (8.14.3/8.14.3/SuSE Linux 0.8) with ESMTP id o46KwMa9131264 for ; Thu, 6 May 2010 15:58:23 -0500 Received: from bombadil.infradead.org (localhost [127.0.0.1]) by cuda.sgi.com (Spam Firewall) with ESMTP id 4DA48155C494 for ; Thu, 6 May 2010 14:00:27 -0700 (PDT) Received: from bombadil.infradead.org (bombadil.infradead.org [18.85.46.34]) by cuda.sgi.com with ESMTP id wjqI5KUYvLCvOomc for ; Thu, 06 May 2010 14:00:27 -0700 (PDT) Date: Thu, 6 May 2010 17:00:26 -0400 From: Christoph Hellwig Subject: Re: [PATCH 04/11] xfs: modify buffer item reference counting V2 Message-ID: <20100506210026.GA30264@infradead.org> References: <1273110351-2333-1-git-send-email-david@fromorbit.com> <1273110351-2333-5-git-send-email-david@fromorbit.com> MIME-Version: 1.0 Content-Disposition: inline In-Reply-To: <1273110351-2333-5-git-send-email-david@fromorbit.com> List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: xfs-bounces@oss.sgi.com Errors-To: xfs-bounces@oss.sgi.com To: Dave Chinner Cc: xfs@oss.sgi.com On Thu, May 06, 2010 at 11:45:44AM +1000, Dave Chinner wrote: > From: Dave Chinner > > The buffer log item reference counts used to take reference?? for every > transaction, similar to the pin counting. This is symmetric (like the > pin/unpin) with respect to transaction completion, but with dleayed logging > becomes assymetric as the pinning becomes assymetric w.r.t. transaction > completion. > > To make both cases the same, allow the buffer pinning to take a reference to > the buffer log item and always drop the reference the transaction has on it > when being unlocked. This is balanced correctly because the unpin operation > always drops a reference to the log item. Hence reference counting becomes > symmetric w.r.t. item pinning as well as w.r.t active transactions and as a > result the reference counting model remain consistent between normal and > delayed logging. > > Signed-off-by: Dave Chinner Looks good and makes the buffer refconting model a lot more sensible. Reviewed-by: Christoph Hellwig _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs