From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1MNm1V-00073l-Jp for qemu-devel@nongnu.org; Mon, 06 Jul 2009 07:06:13 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1MNm1Q-00073Q-KT for qemu-devel@nongnu.org; Mon, 06 Jul 2009 07:06:12 -0400 Received: from [199.232.76.173] (port=43483 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1MNm1Q-00073N-Dq for qemu-devel@nongnu.org; Mon, 06 Jul 2009 07:06:08 -0400 Received: from mx2.redhat.com ([66.187.237.31]:47678) by monty-python.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1MNm1P-00076r-OJ for qemu-devel@nongnu.org; Mon, 06 Jul 2009 07:06:08 -0400 Message-ID: <4A51DA57.3010600@redhat.com> Date: Mon, 06 Jul 2009 13:04:55 +0200 From: Kevin Wolf MIME-Version: 1.0 Subject: Re: [Qemu-devel] [PATCH] qemu-io: add flag to mark files growable References: <20090704155330.GA4825@lst.de> <4A51BBA1.4090307@redhat.com> <20090706102023.GA7835@lst.de> <4A51D486.10005@redhat.com> <20090706105111.GA9469@lst.de> In-Reply-To: <20090706105111.GA9469@lst.de> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Christoph Hellwig Cc: qemu-devel@nongnu.org Christoph Hellwig schrieb: > On Mon, Jul 06, 2009 at 12:40:06PM +0200, Kevin Wolf wrote: >> Ok, you are right. raw was the only exception because it could be used >> as a protocol even though it didn't have a protocol name. >> >> Looks fine to me then if you enforce the limitation instead of only >> printing an error message and opening the image anyway. Not printing the >> error and opening the image would be fine either. I just don't like >> error messages without errors. > > Well, the primary use of qemu-iotests is to try out the I/O patterns, > and the -g option is to check out we do the correct thing for growable > files, so I'd really prefer it to fail if we're not allow to actually > make it growable. Perfectly reasonable. But then make it really fail. > Now one think I could do is to just add a growable flag to the > BlockDriver to make it explicit. Currently only raw files (in the posix > and win32 flavours) would set it. Good point actually, your current logic is wrong: raw-win32 doesn't have the file protocol. I'm not sure though if introducing a new flag just for qemu-io is right. After all qemu-io should test what the drivers provide to qemu and not extend them to work with qemu-io. Currently the drivers don't have the concept of non-growable images. > That would also fix the bug what we > currenly allow to grow host devices even if we can't. How that? The host device drivers don't have a protocol name, do they? Kevin