From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from [140.186.70.92] (port=46909 helo=eggs.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1OCr0p-0003ji-OH
	for qemu-devel@nongnu.org; Fri, 14 May 2010 05:16:56 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.69)
	(envelope-from <armbru@redhat.com>) id 1OCr0n-0001uT-Om
	for qemu-devel@nongnu.org; Fri, 14 May 2010 05:16:55 -0400
Received: from mx1.redhat.com ([209.132.183.28]:16104)
	by eggs.gnu.org with esmtp (Exim 4.69)
	(envelope-from <armbru@redhat.com>) id 1OCr0n-0001uN-Fq
	for qemu-devel@nongnu.org; Fri, 14 May 2010 05:16:53 -0400
From: Markus Armbruster <armbru@redhat.com>
Subject: Re: [Qemu-devel] Re: [PATCH 2/2] Add flush=off parameter to -drive
References: <1273528310-7051-1-git-send-email-agraf@suse.de>
	<1273528310-7051-2-git-send-email-agraf@suse.de>
	<1273528310-7051-3-git-send-email-agraf@suse.de>
	<4BE91708.7050307@redhat.com> <20100511105559.GA21363@lst.de>
Date: Fri, 14 May 2010 11:16:47 +0200
In-Reply-To: <20100511105559.GA21363@lst.de> (Christoph Hellwig's message of
	"Tue, 11 May 2010 12:55:59 +0200")
Message-ID: <m3ocgilu68.fsf@blackfin.pond.sub.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Christoph Hellwig <hch@lst.de>
Cc: Kevin Wolf <kwolf@redhat.com>, Alexander Graf <agraf@suse.de>, qemu-devel@nongnu.org

Christoph Hellwig <hch@lst.de> writes:

[...]
> Markus was looking into separating the block device state in host/
> guest portions lately, and splitting cache= would help with this.
> Driver cache enabled or not (fsync means enabled, O_DSYNC disabled,
> and none probably disabled given that we don't care), while O_DIRECT
> vs normal cached I/O is host state.

Let me provide a bit more context.  -drive is a happy mix of guest and
host parameters.  Some parameters are clearly host, some clearly guest,
and some are... complicated.  For instance, I'd argue that "readonly"
applies to both host and guest.

I'm working on turning all guest parameters proper qdev properties, and
collect the host parameters in a new -blockdev.  -drive stays for
compatibility and convenience (I'd like to deprecate if=none, though).

If you change -drive, please keep the coming separation in mind, and
avoid parameters that mix up guest and host.