From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:49479) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XHwze-0001na-7U for qemu-devel@nongnu.org; Thu, 14 Aug 2014 11:31:16 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XHwzX-0006ze-SK for qemu-devel@nongnu.org; Thu, 14 Aug 2014 11:31:10 -0400 Received: from mx1.redhat.com ([209.132.183.28]:65415) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XHwzX-0006zW-LZ for qemu-devel@nongnu.org; Thu, 14 Aug 2014 11:31:03 -0400 Date: Thu, 14 Aug 2014 17:30:51 +0200 From: Kevin Wolf Message-ID: <20140814153051.GA15821@noname.redhat.com> References: <1406062723-28692-1-git-send-email-mreitz@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1406062723-28692-1-git-send-email-mreitz@redhat.com> Subject: Re: [Qemu-devel] [PATCH v2 0/2] qemu-img: Allow source cache mode specification List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Max Reitz Cc: Peter Lieven , qemu-devel@nongnu.org, Stefan Hajnoczi Am 22.07.2014 um 22:58 hat Max Reitz geschrieben: > Currently, qemu-img does not allow setting the cache mode for source > images. However, it reads images generally only once, therefore a full > writeback cache unnecessarily clutters the host cache. In case the user > finds this undesirable, there has to be a way of disabling that cache. > This series first introduces such a way for check, compare, convert and > rebase and then (while at it) adds a cache mode switch to amend (which > was missing so far). > > Peter already tried to add a BDRV_O_SEQUENTIAL mode to qemu-img convert. > If we want to add such a mode (which in my opinion would be the logical > consequence of this series), we may want to make it just another cache > mode which can be selected by the user. I would prefer if we used -t in all places where only one image is used, so that I don't have to remember that supposedly check -r has only an input image, whereas amend has only an output image. But that's probably a matter of taste... If you don't want to change it, the implementation looks okay: Reviewed-by: Kevin Wolf