From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:57206) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Vcasu-0005Qe-Ct for qemu-devel@nongnu.org; Sat, 02 Nov 2013 09:05:06 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Vcaso-0000RX-Dh for qemu-devel@nongnu.org; Sat, 02 Nov 2013 09:05:00 -0400 Received: from mx1.redhat.com ([209.132.183.28]:57911) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Vcaso-0000Px-2q for qemu-devel@nongnu.org; Sat, 02 Nov 2013 09:04:54 -0400 Message-ID: <5274F86D.8070009@redhat.com> Date: Sat, 02 Nov 2013 14:04:45 +0100 From: Max Reitz MIME-Version: 1.0 References: <1381787553-12497-1-git-send-email-xiawenc@linux.vnet.ibm.com> <1381787553-12497-4-git-send-email-xiawenc@linux.vnet.ibm.com> In-Reply-To: <1381787553-12497-4-git-send-email-xiawenc@linux.vnet.ibm.com> Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH V4 3/6] qcow2: do not free clusters when fail in header update in qcow2_write_snapshots List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Wenchao Xia , qemu-devel@nongnu.org Cc: kwolf@redhat.com, pbonzini@redhat.com, stefanha@gmail.com On 14.10.2013 23:52, Wenchao Xia wrote: > Signed-off-by: Wenchao Xia > --- > block/qcow2-snapshot.c | 7 +++++++ > 1 files changed, 7 insertions(+), 0 deletions(-) > > diff --git a/block/qcow2-snapshot.c b/block/qcow2-snapshot.c > index 4bd494b..c933b7f 100644 > --- a/block/qcow2-snapshot.c > +++ b/block/qcow2-snapshot.c > @@ -304,6 +304,13 @@ static int qcow2_write_snapshots(BlockDriverState *bs, Error **errp) > PRIi64 " with size %" PRIu64 ":%d (%s)", > offsetof(QCowHeader, nb_snapshots), sizeof(header_data), > ret, strerror(-ret)); > + /* > + * If the snapshot data part have been updated on disk, Then the s/have/has/; s/Then/then/ > + * clusters at snapshot_offset may be used in next snapshot operation. > + * If we free those clusters in fail path, they may be allocated and > + * made dirty causing damage, so skip cluster free for safe. s/for/to be/ > + */ > + snapshots_offset = 0; > goto fail; > } Other then that: Reviewed-by: Max Reitz