On Thu, Nov 21, 2019 at 08:38:13AM +0100, Peter Krempa wrote:
> On Wed, Nov 20, 2019 at 18:12:18 +0100, Pavel Mores wrote:
> > On Wed, Nov 20, 2019 at 12:14:11PM +0100, Peter Krempa wrote:
> > > On Wed, Nov 20, 2019 at 11:44:52 +0100, Pavel Mores wrote:
> > > > Since the blockcommit operation is
On Wed, Nov 20, 2019 at 18:12:18 +0100, Pavel Mores wrote:
> On Wed, Nov 20, 2019 at 12:14:11PM +0100, Peter Krempa wrote:
> > On Wed, Nov 20, 2019 at 11:44:52 +0100, Pavel Mores wrote:
> > > Since the blockcommit operation is asynchronous, this has conceptually two
> > > parts. First, we have to
On Wed, Nov 20, 2019 at 12:14:11PM +0100, Peter Krempa wrote:
> On Wed, Nov 20, 2019 at 11:44:52 +0100, Pavel Mores wrote:
> > Since the blockcommit operation is asynchronous, this has conceptually two
> > parts. First, we have to propagate the flag from qemuDomainBlockCommit()
> > (which was
On Wed, Nov 20, 2019 at 11:44:52 +0100, Pavel Mores wrote:
> Since the blockcommit operation is asynchronous, this has conceptually two
> parts. First, we have to propagate the flag from qemuDomainBlockCommit()
> (which was just ignoring it until now) to qemuBlockJobDiskNewCommit(). Then
> it
Since the blockcommit operation is asynchronous, this has conceptually two
parts. First, we have to propagate the flag from qemuDomainBlockCommit()
(which was just ignoring it until now) to qemuBlockJobDiskNewCommit(). Then
it can be stored in the qemuBlockJobCommitData structure which holds