From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Subject: Re: [patch 8/8] fs: add i_op->sync_inode Date: Wed, 29 Dec 2010 10:12:46 -0500 Message-ID: <20101229151246.GA22033@infradead.org> References: <20101218014634.943276411@kernel.dk> <20101218015117.759480620@kernel.dk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, Andrew Morton To: Nick Piggin Return-path: Content-Disposition: inline In-Reply-To: <20101218015117.759480620@kernel.dk> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org - the sync_inodes_sb export removal looks fine to, but should be a separate patch with a good changelog - why is the sync_inode_metadata wait parameter removed? Especially as we would need it for some callers that need to be converted to it. E.g. in this patch converts writeback_inode from non-blocking to blocking behaviour. - except for that the sync_inode() removal is fine, I had planned for that already. But again, please a separate and well-documented patch. As for the actualy sync_inode operation: I don't think it's helpful. The *_sync_inode helpers in ext2 and fat are fine, but there's little point in going through an iops vector for it. Also adding the file syncing really complicates the API for now reason, epecially with the range interface.