From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [RFC] new ->perform_write fop Date: Fri, 21 May 2010 11:15:18 -0400 Message-ID: <20100521151518.GA12752@infradead.org> References: <20100512212403.GE3597@localhost.localdomain> <20100513013926.GD27011@dhcp231-156.rdu.redhat.com> <20100514010042.GI13617@dastard> <20100514033057.GL27011@dhcp231-156.rdu.redhat.com> <20100514064145.GJ13617@dastard> <20100514072219.GC4706@laptop> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Dave Chinner , Josef Bacik , linux-fsdevel@vger.kernel.org, chris.mason@oracle.com, hch@infradead.org, akpm@linux-foundation.org, linux-kernel@vger.kernel.org To: Nick Piggin Return-path: Content-Disposition: inline In-Reply-To: <20100514072219.GC4706@laptop> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org Nick, what exactly is the problem with the reserve + allocate design? In a delalloc filesystem (which is all those that will care about high performance large writes) the write path fundamentally consists of those two operations. Getting rid of the get_blocks mess and replacing it with a dedicated operations vector will simplify things a lot. Punching holes is a rather problematic operation, and as mentioned not actually implemented for most filesystems - just decrementing counters on errors increases the chances that our error handling will actually work massively.