From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: [PATCH 1/3] xen/blk[front|back]: Use the full FLUSH | FUA instead of just FLUSH. Date: Sat, 10 Sep 2011 20:33:06 -0400 Message-ID: <20110911003306.GB9989@oracle.com> References: <1315593060-20031-1-git-send-email-konrad.wilk@oracle.com> <1315593060-20031-2-git-send-email-konrad.wilk@oracle.com> <20110910180849.GA6621@infradead.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20110910180849.GA6621@infradead.org> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Christoph Hellwig Cc: axboe@kernel.dk, xen-devel@lists.xensource.com, linux-kernel@vger.kernel.org, JBeulich@novell.com, jeremy.fitzhardinge@citrix.com List-Id: xen-devel@lists.xenproject.org On Sat, Sep 10, 2011 at 02:08:49PM -0400, Christoph Hellwig wrote: > On Fri, Sep 09, 2011 at 02:30:59PM -0400, Konrad Rzeszutek Wilk wrote: > > During a FLUSH we can pass sector number that we want to > > have flushed - which is what FUA requests are. > > No, that is not the case. > > REQ_FLUSH without data -> pure flush > REQ_FLUSH with data -> preflush plus write Excellent. So we have been doing it right all along. > REQ_FUA -> write and ranged postflush Ah, somehow I was thinking that you can't write data with a REQ_FLUSH, but that is nonsense as the block/blk-flush.c explains in great details. Will drop this patch - and thanks for clarifying it!