From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:58510) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VKZMb-0001dW-C0 for qemu-devel@nongnu.org; Fri, 13 Sep 2013 15:49:15 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VKZMV-0005qR-PD for qemu-devel@nongnu.org; Fri, 13 Sep 2013 15:49:09 -0400 Received: from mx.ipv6.kamp.de ([2a02:248:0:51::16]:33057 helo=mx01.kamp.de) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VKZMV-0005q2-Dh for qemu-devel@nongnu.org; Fri, 13 Sep 2013 15:49:03 -0400 Message-ID: <52336C21.2050902@kamp.de> Date: Fri, 13 Sep 2013 21:48:49 +0200 From: Peter Lieven MIME-Version: 1.0 References: <1379067909-22984-1-git-send-email-pl@kamp.de> <1379067909-22984-9-git-send-email-pl@kamp.de> <5232EA90.7030903@redhat.com> <523358AF.9000504@redhat.com> In-Reply-To: <523358AF.9000504@redhat.com> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH 08/12] qemu-img: conditionally zero out target on convert List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eric Blake Cc: kwolf@redhat.com, ronniesahlberg@gmail.com, stefanha@redhat.com, qemu-devel@nongnu.org, anthony@codemonkey.ws, Paolo Bonzini Am 13.09.2013 20:25, schrieb Eric Blake: > On 09/13/2013 04:36 AM, Paolo Bonzini wrote: >> Il 13/09/2013 12:25, Peter Lieven ha scritto: >>> if the target has_zero_init = 0, but supports efficiently >>> writing zeroes by unmapping we call bdrv_zeroize to >>> avoid fully allocating the target. this currently >>> is designed especially for iscsi. >>> >>> Signed-off-by: Peter Lieven >>> --- >>> qemu-img.c | 22 ++++++++++++++++------ >>> 1 file changed, 16 insertions(+), 6 deletions(-) >>> >>> diff --git a/qemu-img.c b/qemu-img.c >>> index 3e5e388..6eaddc6 100644 >>> --- a/qemu-img.c >>> +++ b/qemu-img.c >>> @@ -1354,7 +1354,8 @@ static int img_convert(int argc, char **argv) >>> } >>> } >>> >>> - flags = BDRV_O_RDWR; >>> + flags = BDRV_O_RDWR | BDRV_O_UNMAP; >> I think this should be a new command-line flag. > I agree - while 'sparse by default' may be reasonable, it is also > feasible to want a mode that guarantees expansion rather than unmapped > or sparse. > Ok, so do you find the proposed -S 0 bei Paolo a good choice? If this is supplied I would go as far as completly setting has_zero_init = 0 also for targets which default to 1. This would guaranteed exspansion and full allocation for all drivers. Peter